Unraveling the Life Cycle of Nyssopsora cedrelae: A Study of Rust Diseases on Aralia elata and Toona sinensis

Rust disease poses a major threat to global agriculture and forestry. It is caused by types of Pucciniales, which often require alternate hosts for their life cycles. Nyssopsora cedrelae was previously identified as a rust pathogen on Toona sinensis in East and Southeast Asia. Although this species had been reported to be autoecious, completing its life cycle solely on T. sinensis, we hypothesized that it has a heteroecious life cycle, requiring an alternate host, since the spermogonial and aecial stages on Aralia elata, a plant native to East Asia, are frequently observed around the same area where N. cedrelae causes rust disease on T. sinensis. Upon collecting rust samples from both A. elata and T. sinensis, we confirmed that the rust species from both tree species exhibited matching internal transcribed spacer (ITS), large subunit (LSU) rDNA, and cytochrome oxidase subunit III (CO3) mtDNA sequences. Through cross-inoculations, we verified that aeciospores from A. elata produced a uredinial stage on T. sinensis. This study is the first report to clarify A. elata as an alternate host for N. cedrelae, thus providing initial evidence that the Nyssopsora species exhibits a heteroecious life cycle.


Introduction
Rust disease, caused by members of the order Pucciniales, poses a severe threat to diverse trees and crops [1,2].This group often demonstrates a complex life cycle, which frequently involves switching between primary and alternate host plants.This heteroecious characteristic plays a central role in the epidemiology of rust diseases [3,4], requiring a comprehensive understanding of all host plants involved.
Toona sinensis (syn.Cedrela sinensis; Meliaceae), also known as the red toon or Chinese mahogany, is a tall deciduous tree that grows up to 20 m in height.This tree is native to East and Southeast Asia and is usually grown to produce high-quality timber, which is ideal for crafting furniture and musical instruments because of its sophisticated reddish colour and durability.In East Asia, its young leaves are utilized as a vegetable as well as for treating several ailments in traditional medicine [5,6].
Aralia elata (Araliaceae), also known as the Korean angelica tree, is a woody plant widely distributed throughout East Asia.This plant is often grown as an ornamental tree because of its unique characteristics, including spiny stems, toothed leaves, and clusters of small white flowers bearing black drupes as fruits.This tree is utilized as a traditional medicinal plant for its pharmacological effects, such as its anti-tumour, anti-inflammatory, and hepatoprotective effects [7].In Korea, its young shoots are harvested in the spring and used in various Korean dishes because of their pleasant aroma and soft texture [8].As of 2021, its production had reached 1383 tons, estimated to be worth 20 billion KRW [9].
Nyssopsora cedrelae is known as a rust pathogen that affects Ailanthus altissima, Toona serrata (=Cedrela serrata), and T. sinensis (=C.sinensis) [10], produces uredinial and telial stages, and has been reported in China, Japan, and Korea [11][12][13].However, its life cycle is not fully understood.Through inoculation experiments using basidiospores obtained from teliospores, Kakishima et al. initially reported that this rust species could complete its life cycle on a single host [14], producing aecia (uredinoid aecia), uredinia, and telia entirely on T. sinensis.However, its aecial stage is not distinctly recognized in nature due to the morphological similarities between the aecial and uredinial stages, and no spermogonium has been found [11].Consequently, the life cycle of N. cedrelae remains unclear.
Rust disease of A. elata is widespread throughout Korea.While the spermogonial and aecial stages of this rust have been observed on A. elata, the other stages of its life cycle remain elusive, leading to the conjecture that this rust may be heteroecious, utilizing different host plants for developing the other life stages.To date, two rust species, Nyssopsora asiatica and Puccinia caricis-araliae (also known as Aecidium araliae), have been reported on A. elata [11,12,15].However, the morphologies and life cycles of these species are quite different.N. asiatica is a microcyclic autoecious species forming only telia on Acanthopanax sciadophylloides, Aralia chinensis, A. cordata, A. elata, A. spinosa, Evodiopanax innovans, Kalopanax innovans, and Merrilliopanax listeri [10,11].The spermogonial and aecial stages that occur on A. elata in Korea are similar to those that occur on P. caricis-araliae in their symptoms but differ in morphology, and these stages are frequently observed in areas where N. cedrelae occurs on T. sinensis.Therefore, we suspected that these stages on A. elata are in fact the spermogonial and aecial stages of N. cedrelae.
This study is the first to report A. elata as an alternate host for N. cedrelae, thus providing initial evidence that the Nyssopsora species exhibits a heteroecious life cycle.In the present study, we comprehensively characterized rust disease samples on A. elata and T. sinensis through morphological and molecular phylogenetic analyses as well as cross-inoculation tests.We aimed to identify the causal agent of rust disease on A. elata in Korea and to clarify the relationships of its spermogonial and aecial stages on A. elata with the uredinial and telial stages of N. cedrelae on T. sinensis.

Sample Collection
Thirty-three rust samples from Aralia elata and Toona sinensis were collected across various locations in Korea.Rust-infected leaves were prepared as dried specimens and preserved at the Kunsan National University (KSNUH) and Korea University (KUS-F) for further processing.In addition, three Japanese specimens of N. cedrelae were provided by the herbarium of the Department of Botany, National Museum of Nature and Science, Tsukuba, Japan (TNS-F), for comparison with the Korean samples.All herbarium specimens used for molecular phylogenetic and morphological analyses in this study are summarized in Table 1.
The ITS, LSU, and CO3 sequences were compared to those of the closest related species in the GenBank database using the Basic Local Alignment Search Tool (BLASTn).The sequences of each marker were aligned using the FFT-NS-2 algorithm method in MAFFT version 7 [18].Phylogenetic trees were constructed using the minimum evolution and maximum likelihood methods based on the Tamura-Nei model in MEGA 7 [19].Statistical support for the branches of the phylogenetic trees was evaluated by the bootstrap method with 1000 replicates.Reference sequences from GenBank used in the phylogenetic analysis are listed in Table 2.

Morphological Analysis
The symptoms and macrostructures of rust-infected specimens were observed under a stereomicroscope (M205C; Leica, Wetzlar, Germany).The micromorphological characteristics were examined and photographed using a differential interference contrast (DIC) light microscope (Axio Imager 2; Carl Zeiss, Oberkochen, Germany).At least 50 rust sori and spores were measured per sample, and their measurements are represented as follows: (minimum-) standard deviation towards the minimum-standard deviation towards the maximum (-maximum) (mean).Scanning electron microscopy (SEM) (S-4800+EDS; Hitachi, Tokyo, Japan) was used for detailed morphological analysis.

Cross-Inoculation Experiments
Cross-inoculation experiments were conducted to demonstrate the pathogenicity of aeciospores from A. elata on T. sinensis.Aeciospores from rust-infected leaves of A. elata (KSNUH1831) were harvested using a spore collector (Tallgrass Solutions Inc., Manhattan, NY, USA) and stored in a refrigerator at 4 • C for an hour.Three healthy T. sinensis plants were inoculated by spraying a suspension of aeciospores in sterile water (1.1 × 10 6 ) onto their leaves.Inoculated plants were then kept in a humid chamber at room temperature (25 • C) for three weeks and monitored for rust-symptom development.Two non-inoculated plants served as controls.

Phylogeny
The ITS and LSU rDNA sequences of rust samples collected from A. elata and T. sinensis were identical.Among the 33 rust samples, slight sequence differences were observed at two sites in the ITS region and one site in the LSU region.BLASTn searches revealed that the Korean and Japanese samples were closest to Nyssopsora altissima from Ailanthus altissima in China.However, there were 17 nucleotide differences from N. altissima in the ITS sequences and a gap in the LSU sequences.In the phylogenetic trees of the concatenated alignment of ITS and LSU sequences (Figure 1), samples from both host plants were consistently grouped with the maximum bootstrapping support value, indicating the robustness of this phylogenetic grouping.The phylogenetic trees revealed two distinct clades within the Nyssopsora species based on their host plants.A clade that includes N. cedrelae shares the same host plants, A. elata and T. sinensis, whereas members of the other clade, including N. echinata (type species of Nyssopsora), originated from various host plants.The CO3 sequences, spanning 649 bp, exhibited no sequence differences across all rust samples.In the phylogenetic tree of the CO3 sequences (Figure 2), samples from both host plants formed a distinct clade that had the highest level of bootstrapping support.Moreover, this clade was distinctly segregated from the Gymnosporangiaceae, Pucciniaceae, and Sphaerophragmiaceae families, further underscoring the unique phylogenetic position of our samples.The CO3 sequences, spanning 649 bp, exhibited no sequence differences across all rust samples.In the phylogenetic tree of the CO3 sequences (Figure 2), samples from both host plants formed a distinct clade that had the highest level of bootstrapping support.Moreover, this clade was distinctly segregated from the Gymnosporangiaceae, Pucciniaceae, and Sphaerophragmiaceae families, further underscoring the unique phylogenetic position of our samples.

Pathogenicity
When healthy T. sinensis leaves were inoculated with aeciospores from A. elata (Figure 5A,B as a control), chlorotic spots began to appear on the leaf surfaces two weeks after inoculation (Figure 5C).The symptoms were similar to those observed in the natural environment.After three weeks, all inoculated plants exhibited more pronounced rust

Pathogenicity
When healthy T. sinensis leaves were inoculated with aeciospores from A. elata (Figure 5A,B as a control), chlorotic spots began to appear on the leaf surfaces two weeks after inoculation (Figure 5C).The symptoms were similar to those observed in the natural environment.After three weeks, all inoculated plants exhibited more pronounced rust symptoms and formed yellow uredinia on their leaf surfaces (Figure 5D-G), from which echinulate urediniospores were produced (Figure 5H), matching the morphological features of N. cedrelae.After five weeks, three of these plants persisted in the uredinial stage without progressing to the telial stage in the experiment.
symptoms and formed yellow uredinia on their leaf surfaces (Figure 5D-G), from which echinulate urediniospores were produced (Figure 5H), matching the morphological features of N. cedrelae.After five weeks, three of these plants persisted in the uredinial stage without progressing to the telial stage in the experiment.

Discussion
In the present study, we uncovered the life cycle of the rust pathogen N. cedrelae.Morphologically, the rust samples on Aralia elata, an essential woody plant in Korean cuisine, were somewhat similar to the characteristics of those from Puccinia caricis-araliae [15], another rust species found on A. elata.They had unique large granules, but the aeciospores were smaller than those of P. caricis-araliae (15.45 × 13.61 µm in Nyssopsora cedrelae versus 21.0 × 18.5 µm in P. caricis-araliae).Further, the type of spermogonia differed between the Korean samples and P. caricis-araliae (type 5 versus type 4).The characteristics of the projections on the teliospore walls of the T. sinensis samples corresponded well with those of N. cedrelae rather than of other Nyssopsora species, even though the teliospores

Discussion
In the present study, we uncovered the life cycle of the rust pathogen N. cedrelae.Morphologically, the rust samples on Aralia elata, an essential woody plant in Korean cuisine, were somewhat similar to the characteristics of those from Puccinia caricis-araliae [15], another rust species found on A. elata.They had unique large granules, but the aeciospores were smaller than those of P. caricis-araliae (15.45 × 13.61 µm in Nyssopsora cedrelae versus 21.0 × 18.5 µm in P. caricis-araliae).Further, the type of spermogonia differed between the Korean samples and P. caricis-araliae (type 5 versus type 4).The characteristics of the projections on the teliospore walls of the T. sinensis samples corresponded well with those of N. cedrelae rather than of other Nyssopsora species, even though the teliospores from the present Korean and Japanese samples were smaller than those previously described (Table 3).The features of urediniospores closely matched those of N. cedrelae.Our phylogenetic study supported the notion that N. cedrelae is a rust pathogen affecting Aralia elata in Korea as well as Toona sinensis in Japan and Korea.Although our samples were morphologically similar to Nyssopsora altissima which has been described from Ailanthus altissima in China [20], they exhibited many sequence differences in the ITS regions.As rust diseases pose a significant risk to forestry and agricultural productivity due to their severe impact on crop yield and quality [22,23], understanding their life cycles, especially their alternate hosts, can potentially enhance disease management strategies [3,4].Our study revealed that N. cedrelae, a rust pathogen associated with T. sinensis, has an alternate host, namely, A. elata.The molecular phylogenetic identity found in the rust species affecting A. elata and T. sinensis provides substantial evidence linking the rust diseases on the two trees.The results of the inoculation test demonstrated that A. elata is a spermogonial and aecial host (alternate host) of N. cedrelae, thereby confirming the heteroecious life cycle of this rust pathogen.Kakishima et al. reported its autoecious life cycle, producing aecia (uredinoid aecia), uredinia, and telia, with basidiospore inoculations obtained from teliospores [14].However, this result may be due to inoculum contamination with urediniospores during basidiospore inoculations because spermogonia were not reported in the inoculations, and uredinoid aecia were produced after basidiospore inoculations.These uredinoid aecia are suspected to present as uredinia after infection with urediniospores.Our results resolve the long-standing enigma that is the life cycle of N. cedrelae, contributing to a better understanding of the epidemiology and dispersion of this pathogen.
Globally, thirteen Nyssopsora species have been reported on various woody plants, including Anacardiaceae, Apiaceae, Araliaceae, and Meliaceae [2,11,[24][25][26].To date, their aecial stage has not been observed, leading to the speculation that they exhibit an autoecious life cycle, either microcyclic (producing only the telial stage) or hemicyclic (producing the uredial and telial stages).Our results indicate the potential presence of alternate hosts in the life cycle of the genus Nyssopsora and provide compelling evidence that supports the hypothesis of Henderson [27] that some Nyssopsora species, including N. cedrelae and N. koelreuteriae, might exhibit a heteroecious life cycle by producing an aecial stage on Apiaceae or Araliaceae.This finding represents not only the first observation of the spermogonial and aecial stages but also the first report of host-alternating in the family Nyssopsoraceae.Our results highlight the benefits of integrating traditional cross-inoculation testing with advanced molecular methods for studying rust pathogens and their complex life cycles.

Conclusions
This study represents a substantial advancement in our understanding of the dynamics of rust diseases affecting two economically valuable trees, A. elata and T. sinensis.We revealed the widespread presence of N. cedrelae on A. elata and elucidated its heteroecious life cycle, alternating between A. elata and T. sinensis.This finding emphasizes the potential threat that N. cedrelae poses to the cultivation and economic value of these two species.The insights gained from the current research are crucial for developing efficient approaches for managing rust diseases on these trees.
indicating the robustness of this phylogenetic grouping.The phylogenetic trees revealed two distinct clades within the Nyssopsora species based on their host plants.A clade that includes N. cedrelae shares the same host plants, A. elata and T. sinensis, whereas members of the other clade, including N. echinata (type species of Nyssopsora), originated from various host plants.

Figure 1 .
Figure 1.Maximum likelihood trees of rust species based on a concatenated alignment of the internal transcribed spacer (ITS) and large subunit (LSU) sequences.Bootstrapping support values (minimum evolution/maximum likelihood) higher than 60% are given above or below the branches.The clade, including Nyssopsora cedrelae, is highlighted in a green box, and the rust samples sequenced in this study are shown in bold.Asterisks (*) indicate sequences of holotype.

Figure 1 .
Figure 1.Maximum likelihood trees of rust species based on a concatenated alignment of the internal transcribed spacer (ITS) and large subunit (LSU) sequences.Bootstrapping support values (minimum evolution/maximum likelihood) higher than 60% are given above or below the branches.The clade, including Nyssopsora cedrelae, is highlighted in a green box, and the rust samples sequenced in this study are shown in bold.Asterisks (*) indicate sequences of holotype.

Figure 2 .
Figure 2. Maximum likelihood trees of rust species based on cytochrome oxidase subunit III (CO3) rDNA sequences.Bootstrapping support values (minimum evolution/maximum likelihood) higher than 60% are given above the branches.The clade, including Nyssopsora cedrelae, is highlighted in a green box, and the rust samples sequenced in this study are shown in bold.

Figure 2 .
Figure 2. Maximum likelihood trees of rust species based on cytochrome oxidase subunit III (CO3) rDNA sequences.Bootstrapping support values (minimum evolution/maximum likelihood) higher than 60% are given above the branches.The clade, including Nyssopsora cedrelae, is highlighted in a green box, and the rust samples sequenced in this study are shown in bold.

Figure 5 .
Figure 5. Cross-inoculation test.(A) Inoculation of aeciospores from Aralia elata on Toona sinensis leaves.(B) Controls.(C) the leaves with chlorotic spots two weeks after inoculation.(D,E) the rust symptoms on the upper (D) and lower (E) surfaces of infected leaves three weeks after inoculation.(F,G) Uredinia on infected leaves by inoculation.(H) Urediniospores from emerging uredinia.

Figure 5 .
Figure 5. Cross-inoculation test.(A) Inoculation of aeciospores from Aralia elata on Toona sinensis leaves.(B) Controls.(C) the leaves with chlorotic spots two weeks after inoculation.(D,E) the rust symptoms on the upper (D) and lower (E) surfaces of infected leaves three weeks after inoculation.(F,G) Uredinia on infected leaves by inoculation.(H) Urediniospores from emerging uredinia.

Table 1 .
Rust specimens of Aralia elata and Toona sinensis used in the present study.

Table 2 .
List of reference sequences of Nyssopsoraceae used for phylogenetic analysis.
a The morphological characteristics were measured from a specimen collected on Cedrela sinensis (now Toona sinensis) (Botanical Garden, Tokyo, Japan, 15 October 1891, S. Hori), without a herbarium number.