Phylogenetic relationships among New Zealand chestnut selections and the relationship with graft failure

Abstract

Graft failure is a major problem in the clonal propagation of New Zealand (NZ) chestnut (Castanea sp.) selections. At present there is no convincing evidence to either support or reject the idea of a specific incompatibility mechanism operation. The NZ chestnut selections are presumed to be hybrids resulting from uncontrolled open pollination of the introduced chestnut species in NZ, most especially, C. sativa (European species) and C. crenata (Japanese species). It is believed that an understanding of the relatedness of the NZ chestnut selections and chestnut species would forster a better understanding of the problems of graft failure in NZ chestnut selections.

To determine the relationships among the NZ chestnut selections and between the NZ chestnut selections and introduced known chestnut species, two different character data sets were generated: (1) random amplified polymorphic DNA (RAPD) based on polymerase chain reaction (PCR) and (2) Morpho-Nut characters.

The first stage of the RAPD studies involved establishing a standard protocol for DNA isolation and RAPD reactions for chestnuts in NZ. Three chestnut species namely; C. sativa (European), C. dentata (American) and C. mollissima (Chinese), were used for this study.

The results showed that RAPDs could be used as a source of genetic markers in chestnut after the protocol was optimised.

The second stage was to determine relationships among chestnut species, the NZ chestnut selections and between the chestnut species and NZ chestnut selections. Twelve accessions of four chestnut species, three F1 interspecific hybrids, and eighteen NZ chestnut selections were used. RAPD data were analysed using the unweighted pair group mean (UPGMA) method of cluster analysis to generate dendrograms. The dendrograms grouped the chestnut species as well as the NZ chestnut selections based on their geographic locations. The South Island selections grouped with the C. sativa species while most of the North Island selections grouped with the C. crenata species. Two North Island selections; 1002 and 1007, which were thought to be hybrids of C. crenata and C. sativa showed strongly C. mollissima-like RAPD patterns.

Data from 31 Morpho-Nut characters using five accessions of chestnut species and eighteen NZ chestnut selections were analysed using Cluster analysis (UPGMA) and Principal Component Analysis (PCA). PCA showed relationships which were similar to those determined by RAPD studies. The C. mollissima-like nature of 1002 and 1007 was independently reinforced.

Experiments to measure graft failure were set up without prior knowledge of the relationships of the NZ chestnut selections. Early and late graft failures were observed which agreed with previous publications in chestnuts. Various degrees of failure were observed which were found to be strongly clonal. Grafting scion wood from one tree on to rootstock grown from nuts of the same clone was found not to be a solution to graft failure contrary to published procedures. Graft failure from this type of grafting was found to be strongly clonal. The North Island selections (except the C. mollissima-like 1002 and 1007) which were C. crenata-like were most strongly implicated in this form of graft failure. The reasons for this graft failure are not certain but use of C. sativa and C. crenata hybrids in either the scion or stock in grafts is strongly implicated as a major cause.