Characterization of the complete plastomes of two flowering epiparasites (Phacellaria glomerata and P. compressa, Santalaceae, Santalales): gene content, organization and plastome degradation CURRENT STATUS: UNDER REVIEW

Backgrounds: The transition to a heterotrophic lifestyle triggers reductive evolution of plastid genome (plastome) in both photosynthetic and non–photosynthetic parasites. A plant parasite parasitizing another plant parasite is referred to as epiparasitism, which is extremely rare in angiosperms. In despite of the particularly special lifeform of epiparasitic plants, their plastomes have not been characterized to date. Sequending such plastomes may enable new insights into the evolutionary pathway of plastome degradation associated with parasitism. Results: In this study, we generated complete plastomes of Phacellaria compressa and P. glomerata (Santalaceae, Santalales) through Illumina shotgun sequencing. Plastome assembly and comparison indicated that plastomes of both species exhibit the quadripartite structure typical of angiosperms, and that they possess similar size, structure, gene content, and arrangement of genes to other hemiparasites in Santalales, especially to those hemiparasites in Santalaceae. The plastomes of P. compressa and P. glomerata were characterized by the functional loss of plastid–encoded NAD(P)H–dehydrogenase and infA genes, which strongly coincides with the general pattern of plastome degradation observed in Santalales hemiparasites. Conclusion: Our study demonstrates that the shift to epiparasitism and reduced vegetative bodies in P. compressa and P. glomerata do not appear to cause any unique plastome degradation compared with their closely related hemiparasites. The epiparasitic lifestyle or an endophytic growth form observed in these two epiparasites may have limited impact on the reductive modification of their plastomes.

3 plastomes have been subjected to strong selective pressures that tend to maintain conservatism in terms of genome size, structure, gene content, and organization [2].
Parasitism in plants is one of the outcomes of the transition from autotrophy to heterotrophy, involving a decrease or complete loss of photosynthetic capacity, and acquisition of nutrition and water from hosts [13]. It is estimated that parasitism has independently evolved at least 11 or 12 times in angiosperms [14,15], and there are approximately 4,500 parasitic plants found in 20 families of flowering plants [13].
A plant parasite parasitizing another plant parasite is termed as epiparasitism [28], hyperparasitism [29], secondary parasitism [30], or double parasitism [31]. This phenomenon is extremely rare in angiosperms, which has only been found in a few species 4 (27 species to date) in Santalales [32]. It is notewortht that all known epiparasites are capable of photosynthesis [13], they are thus regarded as a kind of paricularly special hemiparasitim [29]. Because of the rarity, the epiparasitic plastomes have remained uncharacterized to date.
Among known epiparasitic plants, about one third of the species belong to the genus Phacellaria [32] (Santalaceae, Santalales [33]). Phacellaria is found in the tropical and subtropical areas of Southeast Asia, and consists of eight species that mainly parasitize Loranthaceae hemiparasites [34]. Although all Phacellaria species are capable of photosynthesis, their vegetative bodies grow endophytically in the parasitic host [34].
This growth form in hemiparasites may represent a transitional stage in the evolution toward holoparasitism [13]. Characterization of the plastomes of Phacellaria species may enable insights into the evolution of plastomes associated with the lifestyle shift from hemiparasitism to holoparasitism. In addition, the epiparasitic lifestyle and the reduced vegetative bodies observed in Phacellaria species mean a greater reliance on host plants for nutrients [29]. Given that a previous study suggested that the varying nutritional dependences on host may influence the plastome evolution in hemiparasites [16], the comparison of plastome features and degree of platome degradation with their hemiparasitic relatives can offer information in elucidating whether a higher level of nutritional reliance on the host plants in Phacellaria species leads to a higher level of plastome degradation.
In this study, two epiparasites, namely, Phacellaria compressa and P. glomerata were sampled for Illumina shotgun sequencing. Specifically, we aimed to: (1) assemble and characterize their complete plastomes;(2) compared the plastomes among representatives of Santalales in order to gain insights into the reductive evolution of plastome associated with parasitism; (3) examine whether the epiparasitic lifestyle and the reduced vegetative 5 bodies leads to some unique plastome degradation in both species.

Methods
Plant sampling and genome shotgun sequencing Specimens of P. compressa and P. glomerata (Fig. 1)

Plastome assembly, annotation and comparison
Raw Illumina reads were filtered by NGS QC tool kit [36] to remove adaptors and lowquality reads. The trimmed reads were assembled into contigs with SPAdes v3.10.1 [37].
The representative plastid contigs were extracted and checked using BLAST searches against the reference plastome of Dendrotrophe varians (MF592987). The position and direction of each contig was manually adjusted according to the reference plastome using Geneious V10.2 [38]. The resulting plastomes were annotated using the Dual Organellar Genome Annotator database [39], by carrying out BLAST searches of all plastid genes against the assembled plastomes. Protein-coding genes with one or more frameshift mutations or premature stop codons were regarded as pseudogenes. Start and stop 6 codons and intron/exon boundaries for protein-coding genes were checked manually.

Phylogenetic analysis
In order to investigate the relationships of P. compressa and P. glomerata within Santalales, their plastomes together with all published Santalales plastomes ( Additional file 1) were included in the phylogenomic analysis. Haloxylon persicum (Amaranthaceae, Caryophyllales) was employed as an outgroup. The alignment of plastomes was made using MAFFT [46], and manually edited where necessary. Phylogenetic analyses consisted of standard maximum likelihood (ML) and Bayesian inference (BI) methods. ML analysis was 7 reconstructed using RAxML-HPC BlackBox version 8.1.24 [47] with 1,000 replicates of rapid bootstrap (BS) under the GTRCAT model. The best substitution model (GTR + I + G) for BI analysis was selected using the program Modeltest v3.7 [48] with the Akaike information criterion [49]. BI was performed with MRBAYES v.3.1.2 [50], and the posterior probability values (PP) were run with trees sampling every 100 generations for one million total generations, with the first 25% discarded as burn-in. Stationarity was considered to be reached when the average standard deviation of split frequencies was less than 0.01. Features of the P. compressa and P. glomerata plastomes were presented in Table 1. By comparison with the published plastomes of hemiparasitic plants in Santalales [16,17,19,[41][42][43][44], the size of the complete plastome, LSC, and IR in P. compressa and P.
glomerata was similar to that of Champereia manillana, Dendrotrophe varians and Osyris alba, whilst slightly larger than that of the remaining species in the order ( Table 1). As an obvious SSC reduction was observed in C. manillana, the SSC size of P. compressa and P.
glomerate was closer to that of O. alba and D. varians. The plastomes of P. compressa and P. glomerata had a similar GC content (37.9%), which was unevenly distributed in LSC, SSC, and IRs. The highest GC content was found in the IR regions, followed by LSC. The lowest GC content was observed in the SSC region.
The junctions of IR/SSC and IR/LSC of plastomes in Santalales are divergent (  Fig. 3). The SSC region in Malania oleifera had been extremely contracted, such that no genes are included.
The was recovered. The relationships recovered by our data were highly congruent with previous studies using single or multiple DNA sequences with a more extensive taxon sampling of Santalales, which suggested that root hemiparasitism is the ancestral state in contrast to stem hemiparasitism [52,53].

Discussion
Plastome structure and gene content Santalales consists of 18 families, approximately 160 genera and 2,200 species [53], which comprises almost half of all known parasitical angiosperms, but also includes a relatively small number of autotrophic species in Erythropalaceae, Strombosiacae, and Coulaceae [13,[52][53][54]. It is notable that the reported platomes represent only a small portion of the diversity in the order. Nevertheless, comparison of plastome features among autotrophic, hemiparasitic, and epiparasitic Santalales may provide useful clues for addressing whether epiparasitism influences the plastome evolution in the studied species.
Similar levels of gene loss had been found in other hemiparasitic plants [7,10,18].
Despite P. compressa and P. glomerata exhibiting an epiparastic lifestyle, and their vegetative tissues largely reduced, they likely retained the capacity for photosynthesis 11 because their bodies were overall green [34]. IR expansion and contraction often result in size variations in angiosperm plastomes, leading to the lineage-specific gain or loss of a small number of genes in these regions [55,56]. However, large-scale IR expansions and contractions were only observed in a few autotrophic angiosperms [56]. By comparison, this phenomenon seems to be more common in parasitic plants. For instance, the holoparasite Cytinus hypocistis (Cytinaceae, Malvales) has lost each IR region [25]. Additionally, the IR has been reduced to 1,466 bp in the holoparasite Hydnora visseri (Hydnoraceae, Piperales) [24]. A significant IR contraction to the intergenic spacer between ycf2 and trnL-CAA has been observed in the plastome of a hemiparasite in Santalales, S. jasminodora (Schoepfiaceae) [19]. On the other hand, large IR expansions have been observed in the holoparasite Cynomorium coccineum (Cynomoriaceae, Saxifragales) [57], and hemiparasitic Striga species (Orobanchaceae, Lamiales) [10]. Taken together, these findings suggest that dramatic IR shifts not only occur in holoparasites but also in hemiparasites. However, the similarity of the IR size in P.
compressa and P. glomerata to that of most hemiparasitic species in Santalales implies that the epiparasitic lifestyle of both species has not caused obvious structural changes in the plastomes of P. compressa and P. glomerate, as the result of Progressive Mauve Alignment indicated (Additional file 4).

Plastome degradation in Santalales hemiparasites
The plastid NDH complex mediates photosystem I cyclic electron transport and facilitates chlororespiration in plant cells [58]. The NDH complex is comprised of approximately 30 subunits. Of those, 11 subunits ( ndhA, B, C, D, E, F, G, H, I, J, and K) are encoded by the plastome [59]. Functional and physical loss of these genes is commonly observed in parasitic plants, which is regarded as an early response of plastomes in the evolution of a parasitic lifestyle [60]. Similar to the plastomes of previously studied hemiparasites in Santalales [16,17,19,[41][42][43][44] and other angiosperm orders [7,10,18], both P. compressa and P. glomerata have functionally lost all plastid-encoded ndh genes, suggesting that the NDH complex is the only gene group that has been entirely lost (physically or functionally) from the plastomes of photosynthetic parasites. The results further confirm the assumption that in parasitic plants with photosynthetic capacity, the ndh pathway is not indispensable [10,12,61].
In Santalales parasites, infA is another commonly reduced gene. The functional loss of infA and other housekeeping genes from heterotrophic plastomes is one of the greatest enigmas in heterotrophy-associated plastome degeneration [12]. Nevertheless, the phenomenon has been observed in a wide diversity of autotrophic angiosperms [62]. As a result, infA has been regarded as one of the most mobile plastid genes in angiosperms, which are often transferred to and retained in the nucleus [62][63][64]. Accordingly, we assume that the transfer of infA from the plastome to the nuclear genome may occur in Santalales parasitic plants. 13 Despite the fact M. oleifera is a hemiparasite that maintains photosynthetic capacity, it is surprising to observe that psa, psb, and pet genes have been partially deleted from its plastome. To our knowledge, M. oleifera is the only hemiparasitic plant to date in which critical photosynthetic genes have been partially lost from the plastome. If hemiparasitism represents a transitional step toward the evolution of holoparasitism, this observation in M.
oleifera indicate that the reduction or loss of such genes may occur at a rather early stage.
On the other hand, the partial reduction of these genes from the M. oleifera plastome implies that degeneration of photosynthetic capacity may be a gradual process. This process may have initiated at the hemiparasitic stage and is not likely completed until a holoparasitic lifestyle is achieved [3,12].

Conclusion
Overall, the plastomes of the epiparasites P. compressa and P. glomerata possess similar size, gene content, and arrangement to other hemiparasites in Santalales, particularly to those phylogenetically-related hemiparasites. Our data reveal that plastome degradation in P. compressa and P. glomerata greatly coincides with general trends observed in photosynthetic parasites, that is the functional loss of the ndh pathway [12].

Competing Interests
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding
The authors would like to thank the financial support from the Major Program of National