Characterization of resistance

Inoculation. Two experiments were conducted: the first in March 2018, and the second in May 2018. The old leaves of all the genotypes were cut three days before inoculation in order to inoculate only young leaves by spraying them with a suspension of RCH1 and RFOR2 fresh uredospores at a concentration of 1 × 10⁶ mL^-1. The uredospores were suspended in 96% ethanol (1 g L^-1). The plants were placed in an incubation chamber under the same conditions described for inoculum increase, and in both experiments the plants were kept in a greenhouse at 24 °C average temperature. The experiment unit consisted of one pot with one plant, and five replications for each genotype and for each fungus isolate.

Measurement of resistance. The evaluated variables were the latency period (LP) (days from inoculation to sporulation) and infection type (IT), which refers to the reaction of the host tissue to the pathogen invasion and no-invasion. The plants were inspected daily between days three and 21 after inoculation (dai), and to record the IT, the following classification established by ^{Mueller et al. (2003}) was used: R=resistant, no lesions or very few lesions restricted to a hypersensitivity reaction without sporulation; MR=moderately resistant, very few lesions and production of some uredospores; MS=moderately susceptible, a slightly reduced number of lesions or delayed sporulation; S=susceptible, a considerable number of lesions and amount of sporulation.

Morphological identification of the fungus

Identification at the genus level. Bicellular, pedicelled and horizontally septated teliospores inside non-jelly telia were observed; no peridium in palisade was observed in uredia and telia. According to ^{Cummins and Hiratsuka (1996}), these characteristics correspond to the Puccinia genus.

Identification at the species level. Uredospores of both isolates were yellow in color, globose to ellipsoid, with hyaline and equinulated walls, and some were attached to a short pedicel that emerged from the bottom of the uredia. No germinal pores were observed. Measurements of uredospores of the RCH1 isolate were within range: 18.3-22.3 x 15.13-19.3 µm (mean = 20.06 µm ± 1.64 by 17.157 µm ± 1.41, n = 100). Measurements of uredospores of the RFOR2 isolate were: 18-22.7 x 16.4-19.3 µm (mean = 20.17 µm ± 1.81 by 17.749 µm ± 1.28, n = 100) (Figure 1-F).

Teliospores were found only in the RFOR2 isolate, which were ellipsoid and nailed, and had a rounded, angular and eccentric papillae at the apex. Most of the teliospores observed had no septa and measured 30.1-40.8 x 13.3-17.5 µm (mean = 35.887 µm ± 4.70 by 15.518 µm ± 1.67, n = 100) (Figure 1-G). Septated teliospores measured 42.1-51.2 x 14.5-21.2 µm (mean = 46.526 µm ± 3.86 by 17.544 µm ± 2.7, n = 100) (Figure 1-H). The teliospores wall was smooth and turned dark brown as it approached the apex; its sides measured 1 to 1.8 µm (mean = 1.365 µm) and 3.1-5.3 µm at the apex (mean = 4.7 µm ± 0.77, n = 20).

The morphological characteristics of uredospores and teliospores coincided with the description of P. hemerocallidis provided by ^{Hiratsuka et al. (1992}), ^{Williams-Woodward et al. (2001}) and ^{Hernández et al. (2002}). Uredospores of both isolates were slightly smaller than those described by the authors and, the same as in the other American specimens, no septated teliospores were observed. No alternate host was observed in the identified rust, but according to Hiratsuka et al. (1992), P. hemerocallidis has Type 4 spermatogonia.

Molecular identification. After amplification using ITS5 and ITS4 primers, 618 and 623 bp fragments, approximately, were obtained and then deposited in the Genbank (access numbers FJ897533 and FJ897535). The results of BLASTN confirmed that the RCH1 and RFOR2 isolates corresponded to P. hemerocallidis, since the ITS region showed 99% nucleotic similarity with P. hemerocallidis (access number AF479742).

Characterization of resistance. Based on the latency period and type of infection, the evaluated genotypes were grouped in three categories of resistance to P. hemerocallidis (Table 1). No infection differences were observed between isolates and experiments. The type of infection recorded by genotypes was uniform among isolates.

The H. lilioasphodelus species was susceptible because it initially showed abundant, humid chlorotic and bright small lesions three days after inoculation (dai), which sporulated six dai in the form of small and round bright yellow uredia that grew in extended form and had abundant sporulation on leaf upper surface and backside (10 dai) (Figure 2-A). ‘Radiant Greetings’ and ‘Stella d’Oro’ cultivars showed reddish small necrotic lesions five dai, which indicated a hypersensitivity reaction. However, at 10 and 11 dai, respectively, a yellow initial sporulation was observed on some lesions, so they were classified as moderately resistant (Figure 2-B and C). The H. fulva species showed a reduced number of small chlorotic, humid and bright lesions without sporulation at six dai, so it was classified as resistant (Figure 2-D), the same as ‘Cherry Wine’ cultivar, which had no lesions or sporulation (Figure 2-E).

The hypersensitivity reaction (HR) is an active resistance mechanism that operates after the parasite penetrates. In the broad sense, this mechanism is attributed to the presence of major resistance genes that limit the reaction of a specific parasite genotype, but they actually limit a range of interactions ranging from immunity up to a certain level of susceptibility (^{Jones and Clifford, 1983}).

Figure 2. Types of infection caused by P. hemerocallidis in five Hemerocallis spp. genotypes 21 days after inoculation. A) Uredia and uredospores on the H. lilioasphodelus susceptible genotype (10 dai). B) Necrotic lesions (hypersensitivity reaction) and initial sporulation in the moderately resistant ‘Radiant Greetings’ cultivar (10 dai). C) Hypersensitivity reaction and initial sporulation in the moderately resistant ‘Stella d’Oro’ cultivar (11 dai). D) Tiny chlorotic lesions without sporulation in the H. fulva resistant genotype (6 dai); E) No lesions on the ‘Cherry Wine’ resistant cultivar. 

The genes that confer HR cause the host cell to collapse after the haustorium is formed. The cell collapses a few hours after penetration in order to interfere with the fungus nutriment supply, causing a rapid death of the infection hyphae. At first glance, this phenomenon can be observed in the form of yellowish and necrotic spots. In other cases, the collapse is slower and allows the formation of uredospores (^{Parlevliet and van Ommeren, 1975}; ^{Niks et al., 2011}). This explains the type of infection observed in ‘Radiant Greetings’ and ‘Stella d’Oro’ cultivars, because in both cases, hypersensitivity lesions were first observed, but later uredia were detected in some lesions. The studies conducted by ^{Mueller et al. (2003}), ^{Li et al. (2007}) and ^{Buck (2013}) indicate that ‘Stella d’Oro’ cultivar was moderately resistant, resistant and highly resistant, respectively, where the type of infection was limited to necrotic lesions. The reactions observed in the experiments conducted in this study for the cultivar are similar to those reported by the authors, so it seems that there is no variation in the pathogen virulence.

The H. fulva genotype showed slow sporulation on some of the chlorotic lesions outside of the interval of the resistance evaluation, approximately 28 days after inoculation. ^{Parlevliet and van Ommeren (1975}) indicated that genotypes with partial resistance have a susceptible type of infection (complete compatibility between the pathogen and its host, which indicates the absence of hypersensitivity), and that the latency period is far longer, while the infection is less frequent compared to a susceptible host. From the established experiments, major differences can be observed in the latency period and infection frequency (not quantified), but no histological studies, as the studies conducted for cereal rusts (^{Niks, 1983}; ^{Moldenhauer et al., 2006}), were conducted to specify the level of reaction and thus define the existence of a possible partial resistance. Based only on the coloration changes in the host’s tissue, the types of infection in ‘Radiant Greetings’ and ‘Stella d’Oro’ cultivars have only one type of hypersensitivity resistance. However, H. fulva, which showed only one latency period of approximately 28 days and null macroscopically hypersensitivity reaction, could have partial resistance sensu Parlevliet (Parlevliet and van Ommeren (1975), but this must be further studied.

The RCH1 and RFOR2 isolates inoculated in the different genotypes caused the same resistance reactions, which indicates that there are no variations in pathogenicity. However, studies conducted by ^{Hernández et al. (2002}) and ^{Chatasiri et al. (2006}) indicate that there is proof of variation in the pathogen, because by comparing the ITS region of several isolates coming from different America and Asia geographical areas, they observed genetic differences in the specimens from each continent. However, they point out that a wider sampling is needed in order to determine if there is genetic variability in the fungus.

Inoculation of different P. hemerocallidis isolates used in the study conducted by ^{Buck (2013}) confirms the presence of pathotypes with a different type of virulence in daylily cultivars in the southern area of the United States. ^{Carvalho et al. (2018}) report that the absence of the isolate found in Portugal is completely aligned with the sequences from the United States and Costa Rica, but that it differs from the reported sequences from Russia, Japan, Australia and Mexico (sequences reported in the present study).

Rusts are obliged parasites and highly specific to the host specificity; in some rust species the special forms and physiological races are defined according to the host specificity and cultivars (^{Li et al., 2007}). In the case of P. hemerocallidis, there are no reports on its host specificity, so it would not be surprising to find such pathogenic specialization in this pathosystem, considering that it is similar to those that involve rusts in different economically important crops (^{Leonard and Szabo, 2005}; ^{Soto-Estrada et al., 2005}; ^{Yáñez et al., 2009}), but further studies must be conducted to demonstrate this aspect.

It is known that in macrocyclic rusts, including wheat stem rust| (P. graminis f. sp. tritici) and leaf rust (P. triticina), the genetic recombination occurs in the alternate host and that this is a mechanism to create new virulence combinations. Since Patrinia villosa, an alternate P. hemerocallidis host (^{Ono, 2003}), has not been found in Mexico, its possible influence on the development of genetic variation in the fungus is unknown.

In North America, the first report on this rust was submitted by ^{Williams-Woodward et al. in 2001}, when they detected infected material in the United States. Later, the disease quickly spread across America and it is probably present in all the countries where Hemerocallis spp. susceptible cultivars are grown.