Mitochondrial diversity in linguistic isolates of the Alps: A reappraisal

Brief Communication

Abstract In Stenico et al. (1996) we reported unusually high levels of mitochondrial diversity in the Alps. In particular, two communities of Ladin speakers appeared the most extreme European mitochondrial outliers at that time. Recently, it has been observed that some rare nucleotide substitutions occur repeatedly among those sequences, raising the possibility of systematic sequencing errors. No biological material was left from the previous study, and hence we had to sample new individuals from the same communities. Here, we present the HVSI sequence variation, along with haplogroup assignment based on restriction fragment length polymorphism (RFLP), in 20 Ladin speakers of Colle Santa Lucia. None of the new sequences displays substitutions at the sites viewed as problematic. However, Ladins still show high levels of mtDNA diversity, both within their community and with respect to other Europeans, and they can still be considered one of the main European mitochondrial outliers.

KEY WORDS: ALPS, HVSI SEQUENCE, LADIN SPEAKERS, MTDNA DIVERSITY

In a previous study (Stenico et al. 1996), we described unusually high levels of mitochondrial DNA (mtDNA) diversity in seven communities of the Alps. With the exception of one group, the Mocheni, whose sequences suggested a recent founder effect, variation of the hypervariable segment I, HVSI, was high both within and among populations, so much so that a language isolate, the Ladins, appeared at that time the most extreme European mitochondrial outlier. We interpreted those results as suggesting that, despite their current small number (30,000 to 35,000 individuals: http://www.ethnologue.com/show_language.asp?code= LLD), Ladin speakers are descended from an originally large and genetically variable population, which had been isolated from its neighbors of the plains. In particular, a high frequency of an elsewhere-rare haplogroup, T (Richards et al. 1996), was apparent among Ladins. The presence of a T motif [16126, 16292, 16294, according to the numbering in Anderson et al. (1981)] in an individual who lived in the same area 6000 years ago suggests some degree of genealogic continuity, at least from the earliest Neolithic period (Di Benedetto et al. 2000).

Recently, Hans-Jurgen Bandelt (personal communication) informed us that nucleotide substitutions at six sites, 16085, 16106, 16221, 16313, 16315, and 16246, occur repeatedly in the sequences of Stenico et al. (1996), although they are rarely seen in other European samples, and never in the same branches of the mitochondrial network. He attributed those mutations ('phantom' mutations; Bandelt et al. 2001) to systematic sequencing errors.

The DNA extraction and typing for Stenico et al. (1996) was carried out in Padua, in the laboratory headed by Loredana Nigro, who passed away in 1998. Recovering the original electropherograms, or the blood samples for resequencing, proved impossible. We then decided to collect new samples from the same communities. Here, we report on the HVSI sequences from 20 Ladin speakers (from all of whom informed consent was obtained) of Colle Santa Lucia, whose maternal grandmothers had been local residents, and who were not aware of any consanguinity along their female lines. Given the increasing diffusion of the Italian language and the small census size (480 individuals in 1991; www.dossier. net/pop.htm), collecting larger numbers of unrelated Ladins in that community appears hardly possible.

Materials and Methods

DNA Analysis. DNA was extracted from 200 (mu)L of fresh blood with the Genomix kit (Amersham Pharmacia) following the manufacturer's instructions. To avoid the occurrence of mutations that appear only in some batches of polymerase chain reaction (PCR) and sequencing reactions, we amplified HVSI at least twice in each individual (for PCR conditions and primers see Stenico et al. 1998), each time from an independent DNA extract from whole fresh blood. We then performed several cycle-sequencing reactions (Big Dye Terminator kit, Applied Biosystems) of both the L and the H strand, from independent amplicons. In addition, all individuals were typed for specific restriction polymorphic sites, which define haplogroups M, L3, HV, H, J, T, pre-JT, I, and UK, according to Torroni et al. (1996) and Macaulay et al. (1999).

Data Analysis. Indices of standard and molecular diversity and genetic distances, d^sub ab^ (Nei 1987), were calculated with ARLEQUIN (ver. 2.00; Schneider et al. 2000). Negative d^sub ab^ values, which occur when within-population diversity exceeds the diversity between two populations, were corrected by adding a constant quantity to each value in the matrix. A neighbor-joining tree (Saitou and Nei 1987) was then constructed by means of PHYLIP (ver. 3.57c; Felsenstein 1995).

The average genetic distance, d^sub E-L^, between Ladins and the European populations in Stenico et al. (1996) was compared to the average genetic distance between random pairs of the same European populations, d^sub E-E^. Individuals were randomly extracted with replacement from each population for 1000 times, and d^sub L-E^ and d^sub E-E^ were computed and compared each time.