Seven New Complete Plastome Sequences Reveal Rampant Independent Loss of the ndh Gene Family across Orchids and Associated Instability of the Inverted Repeat/Small Single-Copy Region Boundaries

Abstract

Earlier research has revealed that the ndh loci have been pseudogenized, truncated, or deleted from most orchid plastomes sequenced to date, including in all available plastomes of the two most species-rich subfamilies, Orchidoideae and Epidendroideae. This study sought to resolve deeper-level phylogenetic relationships among major orchid groups and to refine the history of gene loss in the ndh loci across orchids. The complete plastomes of seven orchids, Oncidium sphacelatum (Epidendroideae), Masdevallia coccinea (Epidendroideae), Sobralia callosa (Epidendroideae), Sobralia aff. bouchei (Epidendroideae), Elleanthus sodiroi (Epidendroideae), Paphiopedilum armeniacum (Cypripedioideae), and Phragmipedium longifolium (Cypripedioideae) were sequenced and analyzed in conjunction with all other available orchid and monocot plastomes. Most ndh loci were found to be pseudogenized or lost in Oncidium, Paphiopedilum and Phragmipedium, but surprisingly, all ndh loci were found to retain full, intact reading frames in Sobralia, Elleanthus and Masdevallia. Character mapping suggests that the ndh genes were present in the common ancestor of orchids but have experienced independent, significant losses at least eight times across four subfamilies. In addition, ndhF gene loss was correlated with shifts in the position of the junction of the inverted repeat (IR) and small single-copy (SSC) regions. The Orchidaceae have unprecedented levels of homoplasy in ndh gene presence/absence, which may be correlated in part with the unusual life history of orchids. These results also suggest that ndhF plays a role in IR/SSC junction stability.

Fig 1. Maximum likelihood tree of monocots generated from 79 plastid coding gene sequences (based on both partitioned and unpartitioned data), with support values indicated.

An asterisk on a branch means that it was supported by 100% bootstrap value (BP) in the RAxML trees and a Bayesian posterior probability of 1.00.

Fig 2. Map of the seven newly sequenced orchid plastomes.

Genes are distinguished by colors based on their function. An asterisk indicates a gene with an intron. The direction of transcription is clockwise for genes on the inside of the circle, and counterclockwise for genes on the outside.

Fig 3. ndh gene length variation among orchid plastomes.

Numbers within cells refer to gene or pseudogene lengths. Key to colored cells: white—full-length, in frame; yellow—pseudogenized (very short, frame-shift inducing indels); blue—truncated (significant indels); red—completely deleted. The lengths of ndhA and ndhB include their introns. Taxa are presented in the same order as in Fig 4. Gene lengths are not provided for Apostasia wallichii because in some cases, small portions of the 5′ or 3′ ends of the genes were not sequenced and were hence unavailable in GenBank; likewise, only coding sequences were available (no introns were included) for the two intron-containing ndh loci [19].

Fig 4. Character reconstructions of ndh gene loss and pseudogenization events, mapped onto the Orchidaceae portion of the ML tree from Fig 1.

The insets show graphical representations of the step matrices used in the character reconstructions; see text for details. Taxa are presented in the same order as in Fig 3.

Fig 5. Comparison of IR boundaries among 37 orchid plastomes and Asparagaceae outgroups.

The complete Apostasia plastome is unpublished and hence unavailable for comparison, and the IRs in the two complete Epipogium plastomes are highly truncated along with the rest of the plastome (total plastome length of 30.7 kb in E. aphyllum and 19.0 kb in E. roseum), and thus are not shown here.