Overcoming paralogy and incomplete lineage sorting to detect a phylogeographic signal: a GapC study of Armeria pungens

Abstract

Low-copy nuclear genes have been suggested as a promising source of independent phylogeographic markers in plants. However, the available studies at the intraspecific level have revealed that extracting information from them is frequently hampered by paralogy and lack of coalescence of alleles. It is thus relevant to test their utility with plants for which solid data from other markers are available. The aims of this study are to retrieve phylogeographic useful information in a low-copy nuclear gene by examining the congruence of the genetic variation with the geography, as well as with previous nuclear ribosomal, plastid, and amplified fragment length polymorphism (AFLP) markers. Seven combinations of primers have been assayed to characterize the structure of GapC (cytosolic glyceraldehyde 3-phosphate dehydrogenase) in Armeria pungens (Link) Hoffmanns. & Link, a linearly distributed Atlantic–Mediterranean disjunct sand-dune species. A matrix of 101 direct sequences from 71 individuals was analysed with statistical parsimony. To check the reliability of direct sequencing, 216 cloned sequences were also generated. Tests of recombination have also been attempted. By comparing nucleotide and amino acid sequences, three different paralogs (1, 2, 3) were identified of which paralog 2 was sampled for phylogeographic inference. Within this paralog, 13 alleles belonging in three different sequence types (I, II, III) were detected. These types are shown to correspond with lineages from the same locus whose splitting predates the origin of A. pungens, although type III could be a recent paralog. Allelic variation within types I and II followed a clear geographic trend supporting the two main genetic lineages detected in A. pungens with previous markers. This study suggests that information on the population history of a species can be retrieved, even if some uncertainty remains on the source of variation of low-copy nuclear gene sequences, either alleles from the same locus or paralogs.