DNA polymorphism underlying allozyme variation at a malic enzyme locus (mMEP‐2*) in Atlantic salmon (Salmo salar L.)

Abstract A non‐synonymous single nucleotide polymorphism (SNP) underlies a diallelic allozyme polymorphism at the mitochondrial NADP‐dependent mMEP‐2* locus in Atlantic salmon (Salmo salar L.). The resultant amino acid substitution, which alters the charge of the allelic products, matches the differential mobility of the two allozyme alleles, whereas allozyme and SNP assays revealed genotyping concordance in 257 of 258 individuals. A single mismatch, homozygous allozyme vs. heterozygote SNP, suggests the presence of a second, less common null allele.

A number of population-based studies have highlighted mMEP-2* allele and genotype frequencies that are indicative of a selection signal (e.g. Gilbey et al., 1999;Jordan et al., 1990Jordan et al., , 1997Moran et al., 1998;Verspoor & Jordan, 1989). Further research has stalled, however, with population genetic studies now using direct DNA marker screening platforms. To prime further research into mMEP-2* the authors report on the identification of a causal SNP (accession ss9410532730) for the observed mMEP-2* allozyme polymorphism and the development of simple rapid DNA-based assays to survey it.
Frozen skeletal muscle tissues for both allozyme and DNA screening were opportunistically acquired from archived materials, collected over a 19-year period (1997-2015) from a range of unrelated projects. These comprised 258 individuals (including pedigree parents detailed later) from six different sources: two rivers (Dee and Tay, Scotland n = 36 and 32, respectively), three commercial European farm strains (n = 50, 40 and 40) and a ranch strain (Burrischoole, Ireland n = 60). In addition, archived DNA samples from wild populations inhabiting the latitudinal limits of the species range were also surveyed: two rivers from Finnmark, Norway (n = 46 each; Kongsfjordelva and Repparfjordelva) and two rivers from northern Spain (n = 30 each; Rio Ulla and Rio Bidasoa). Sampling was conducted according to national regulations in place at the time the specimens were taken.
The gene search focused on archived DNAs from two S. salar mapping panels (Br5 and Br6) originally generated from an EU-funded linkage mapping project (SALMAP, 1997(SALMAP, -2000. The salmon families interrogated were outcrosses involving four wild-caught adults (River Tay, Scotland), each pedigree comprising sire, dam and 48 progeny. Three allozyme loci, including mMEP-2*, were found among the c. 350 markers, mostly short tandem repeats (STRs), assigned to a low-resolution map based on these pedigrees (Danzmann et al., 2005).
Using malic enzyme vertebrate homologue sequences (mouse and zebrafish) as a starting point, NCBI and EBI databases (plus salmonid TIGR and GRASP EST repositoriesno longer accessible) were interrogated to identify potentially relevant mRNA sequences in S. salar.  Table S1 for details). Independent segregation (P = 0.8) was found between the LOC106581960 STR and mMEP-2*. In contrast, there was complete co-segregation between LOC106586750 STR alleles and mMEP-2* alleles (i.e., zero recombinants among the 96 progeny), identifying LOC106586750 as the mMEP-2* gene, whereas LOC106581960 is likely to be mMEP-1*.  (Table 1). The SNP "A" base disrupts the restriction enzyme site (no cutting), whereas the "G" base permits restriction, yielding 1004 and 177 bp products (Figure 1b). A fluorescent allelespecific PCR assay (KASP; LGC Genomics) was also designed as a rapid screening tool (Figure 1c; Table 2).
In total 258 individuals from six different populations/stocks were screened for both allozyme and SNP variability (identifying 61, 124 and 73 as *100/*100, *100/*125 and *125/*125 allozyme "genotypes," respectively). Polymorphism was observed in all six stocks. There was almost complete concordance between allozyme and SNP scoring (257 of 258 individuals). A single fish was scored as *100/*100 for allozyme and AG with both RFLP and KASP assay.
The entire CDS of this individual was sequenced, confirming the AG genotype. No other sequence differences were found. A null allele, caused by mutation in a regulatory gene region, could account for this observation. With genomic resources for species expanding at a rapid rate, the identification of DNA mutations underlying allozyme variation is becoming more straightforward. In the current study, the use of mapping pedigrees to confidently distinguish between duplicate loci was particularly helpful. The sole point mutation within the CDS of alternate allozyme homozygote sibs, causing a charge changing amino acid substitution matching electrophoretic expectations, is compelling evidence that this is a causal SNP. Furthermore, there was extremely high concurrence between allozyme and SNP assays from the same individuals (>99.5%), and similar allele frequency disparity was observed for both SNP assay (this study) and allozyme assay (Verspoor et al., 2005) among populations from extremes of the S. salar latitudinal range. Both observations lend support for the SNP polymorphism being a robust proxy for previously reported mMEP-2* allozyme polymorphism.
A single mismatch, homozygous allozyme vs. heterozygote SNP genotype, was observed. This could be indicative of an additional low frequency "null" allele, which could potentially generate a false selection signal in samples, if present at a higher frequency. Nonetheless, a null allele, which should lead to an overestimation of homozygote numbers, would not clearly explain the excess of allozyme heterozygotes found among grilse reported by Jordan et al. (1990). Further work is needed to clarify this issue. There is a large array of potential regulatory mechanisms that can underlie null alleles (Rojano et al., 2019), which can be extremely difficult to identify solely from DNA sequence data (e.g., Saha et al., 2022).
This characterised DNA polymorphism and assay will allow for more intensive work into the widely suspected selective action at mMEP-2*. As well as the single locus assays described above, it should be straightforward to include this assay into existing bespoke SNP panels used routinely for population and pedigree screening, and it is already represented multiple times on an existing Atlantic salmon high-density genotyping array (Houston et al., 2014).

AUTHOR CONTRIBUTIONS
J.B.T.: conception, data generation, data analysis and manuscript preparation. M.J.L. and M.B: data analysis and manuscript preparation.
T A B L E 2 KASP allele-specific primer details