Introduction

Rusts are important plant diseases caused by pathogens belonging to the Pucciniales order, which is the most extensive taxonomic order of plant pathogenic fungi, encompassing over 8000 species (Toome-Heller 2016). Widely distributed, rust-causing agents have specialized across various hosts and climates, being obligate basidiomycete pathogens of annual crops, shrubs, and trees worldwide (Helfer 2013). Due to its negative impact on crop** systems, two types of rust have been included among the 10 plant fungal pathogens of greatest scientific and economic relevance (Dean et al. 2012). These belongs to Puccinia spp., responsible for rust in cereals and some legumes, and Melampsora lini, which causes rust in flax (Dean et al. 2012). Although not included in this ranking, soybean rust caused by Phakopsora pachyrhizi warrants special attention due to its recent surge in incidence globally, particularly in regions where soybean is a main crop (Dean et al. 2012; Goellner et al. 2010).

Most rust species are macrocyclic heteroecious fungi. They present complex life cycles involving various spore types, that infect different host species (Duplessis et al. 2021). Their life cycle combines sexual and asexual stages which renders them high-risk evolutionary pathogens, capable of overcoming plant defences with relative ease (Mapuranga et al. 2022). The life cycle stages of rusts are traditionally referred to by Roman numerals:

  • Pycniospores (Stage 0): Produced in pycnidia, these serve as haploid gametes in heterothallic rusts.

  • Aeciospores (Stage I): Arising from aecia, these non-repeating, dikaryotic asexual spores infect the primary host.

  • Urediniospores (Stage II): Formed in uredia, these repeating, dikaryotic vegetative spores can cause autoinfection on the primary host. They are often visible as rust-coloured pustules on the plant.

  • Teliospores (Stage III): Produced in telia, these spores typically represent the overwintering stage and lead to the production of basidia and basidiospores.

  • Basidiospores (Stage IV): Arising from teliospores, these wind-dispersed haploid spores typically infect an alternate host, playing a crucial role in the pathogen lifecycle.

Depending on the rust species, the epidemic cycle may be caused by aeciospores (Stage I), or by urediniospores (Stage II) (Singh et al. 2023; Beniwal et al. 2022). In both cases, symptoms are characterized by numerous small, rust-like pustules ranging from orange/yellow to brown that form on infected plant tissues (Fig. 1). These pathogens extract nutrients from infected plant cells through specialized structures called haustoria (Voegele and Mendgen 2003). In some cases, attempt of host defence can be observed as light-yellow halo or dark necrotic area surrounding rust pustules. The disease severity leads to a loss of photosynthetic area of infected tissues (leaves, stems and pods) (Fig. 1). These symptoms result in a decline in grain quality and a reduction in the overall plant yield that vary between 20 and 80% depending on the crop, the cultivar and the environment (Emeran et al. 2011; Childs et al. 2018; Das et al. 2019). Yield losses can even reach 100% under highly favourable condition in terms of air humidity and temperature on susceptible cultivar if the disease is not controlled (Newcombe 2004; Das et al. 2019; Gautam et al. 2022a). This yield penalty cost around 6 M USD annually for soybean cultivation alone in the United States and Canada while the economic losses associated with the pandemic rust infection of soybean in South America rounded the 1.3 billion USD in 2010 (Yorinori 2021; Crop Protection Network 2023).

Fig. 1
figure 1

Illustration of rust symptoms on chickpea and faba bean. The pictures a and b show pustules of Uromyces ciceris-arietini on the adaxial surface of chickpea leaves under controlled (a) or field conditions (b). Image c shows the development of rust pustules incited by the fungus U. viciae-fabae on leaves, stem, and pod of faba bean under field conditions

Full size image

The Fabaceae family, second in global agricultural importance after the Poaceae is fundamental in the context of food security and environment sustainability (Graham and Vance 2003). Legume crops like bean, lentil, alfalfa, and pea constitute 27% of the world primary crop production (Vance et al. 2000). These grain and forage legume crops, that cover 12–15% of arable land, are indispensable in various agronomic systems (Azooz and Ahmad 2015; FAOSTAT 2022). They are particularly crucial in low-income and develo** countries, serving as the main source of grain and fodder for both human consumption and livestock feeding (Mitchell et al. 2022). Legumes are notable for their nutritional richness, providing essential plant-based proteins, vitamins, and minerals, critical for diets worldwide, especially for smallholders and subsistence farmers, as most legumes are recognized as low-input crops (Jha and Warkentin 2020; Venkidasamy et al. 2019; Didinger and Thompson 2021). Additionally, legumes offer environmental benefits, through atmospheric nitrogen fixation, that improves soil structure and benefits subsequent annual crops in rotation (Gungaabayar et al. 2023). Therefore, the production of grain and forage legumes continues to grow globally (FAOSTAT 2022), primarily to meet the high demands for livestock feed for meat and dairy production and to a lesser extent for human consumption due to new plant-based dietary habits (Alexandratos 2012; Rubiales et al. 2021). However, global legume production faces important challenges due to environmental adaptability issues and susceptibility to pest and diseases, with rust disease being a major agent of these problems (Rubiales et al. 2015). These factors hinder the capacity of legume production to meet the growing demands posed by demographic growth, emphasizing the need to improve agricultural practices and disease management to boost legume cultivation.

Given the important losses induced by rust disease on legume annually, improving rust management in these crops is a priority. This air-borne pathogen is difficult to eradicate as it can survive in the field for multiple seasons in a latent stage or by completing its lifecycle on alternative hosts. Therefore, controlling rust can only be addressed by integrating various disease management methods including agricultural practices, biological and chemical control, and breeding. Among these approaches, the use of rust resistant varieties has been highlighted as the most cost-effective, efficient, and environmentally friendly method to prevent the massive losses caused by rusts (Rubiales et al. 2015; Barilli et al. 2014). Here we review the efforts made toward legume rust management including the most recent advances in legume breeding for resistance to rust. This encompasses a range of strategies, from cultural practices to conventional and advanced breeding techniques, reflecting the importance and complexity of combating rust in these crops.

Rust diversity in legumes

Understanding pathogen population and their diversity is essential to design efficient disease management strategies, and to develop new rust resistant cultivars (Sillero et al. 2006). Legume rusts are mainly incited by fourteen fungal species (Table 1). Most rust species belong to the Uromyces genus, although some species from the Phakopsora and Puccinia genus can also be of importance for some legumes. Collectively, these rust species can exhibit a wide and overlap** host range with several Uromyces species able to infect the same host (Zhang et al. 2011). Most rust species have been further classified in different races and/or pathotypes according to their virulence pattern on different host genotypes (Table 1).

Table 1 Causal agents of legume rust diseases
Full size table

Race-specific resistance against P. pachyrhizi has been described in soybean, although comprehensive studies on the physiological races of the pathogen are still needed (Chander et al. 2019). Studies on the population of P. pachyrhizi identified six races on soybean while four distinct races were also detected on kudzu (Pueraria lobata), its wild alternative host (Yamaoka et al. 2002, 2014). In addition, preliminary studies have indicated that P. pachyrizi population from South Africa and Brazil belonged to at least two distinct races (Caldwell and McLaren 2004; Darben et al. 2020). These findings suggest a geographic and host-based pathogenic variation of P. pachyrhizi, highlighting the need for further research to understand the extent and implications of this diversity (Akamatsu et al. 2017).

The existence of different races have also been suggested for the peanut rust pathogen P. arachidis, as susceptibility has been reported on plants presumed to be resistant under tropical climatic conditions (Kuo et al. 2021). However, physiological identification of these different races remains unclear so far (Subrahmanyam et al. 1993; Waliyar et al. 1993). The inability to physiologically differentiate these pathogenic races underscores the challenges in managing peanut rust and suggests a need for ongoing research, particularly in understanding how climatic conditions influence pathogen virulence and host resistance (Kuo et al. 2021). In addition, they highlight the complexity of pathogen-plant interactions and the potential impact of environmental factors on disease dynamics.

A wide global diversity has also been demonstrated for some Uromyces spp. populations, such as U. appendiculatus (syn. U. phaseoli), for which hundreds of races and pathotypes have been described in different regions of the world (Acevedo et al. 2013; Nyang et al. 2016; Liebenberg and Pretorius 2011). Research on the broad range of virulence on common bean genotypes has shown that U. appendiculatus races segregate into two distinct groups, Andean, and Mesoamerican (Pastor-Corrales and Aime 2004). Some of these pathotypes are limited to their geographic origin, such as the Andean races that usually infect common beans from Andean origin (Sandlin et al. 1999). By contrast, races from the Mesoamerican gene pool present a broader range of virulence being able to infect common beans of Andean, Middle American, and Mesoamerican origin (Pastor-Corrales 2004).

Two races have been described within U. phaseoli var. vignae affecting cowpea (Gay 1971). Resistance to race 1 act independently of leaf age. By contrast, resistance to race 2 have been shown to act differentially depending on leaf age, infection site, and cultivar (Heath 1994) suggesting the greatest specificity of cowpea rust from race 2.

Although the existence of races has not been clarified within U. pisi population, it is recognized as the causal agent with the broadest host range among legumes. It significantly affects pea, grass pea and to a lesser extent lentil and vetch (Barilli et al. 2012a, b). Although differences in the level of partial resistance expressed by specific host genotypes have been detected in response to different U. pisi isolates, differential patterns of the hypersensitive response (HR) were not detected impeding the definition of physiological races (Osuna-Caballero et al. 2024). Interestingly, host range studies and phylogenetic analysis of Uromyces species suggested that the other legume infecting Uromyces species might have evolved from U. pisi (Emeran et al. 2008; Chung et al. 2004).

U. viciae-fabae can also affect a wide range of legumes, including faba bean, lentil, pea, and vetch (Conner 1982a; Gautam et al. 2022b). It is a complex species for which host specialization has been suggested (Emeran et al. 2005). Existence of races have been proposed within the faba bean infecting isolates, although not systematically monitored. Up to 9 races have been reported in Australia alone (Ijaz et al. 2020, 2021a), while up to 16 races have been additionally described from isolates of worldwide distribution (Emeran et al. 2001). This classification is based on the presence or absence of necrosis as a criterion for discrimination in broad bean cultivars. Some studies had also postulated existence of races based on pustule size (Conner 1982b; Rashid 1984). Recent studies have demonstrated differences in the pattern of HR expressed by Lens spp. genotypes in response to U. viciae-fabae isolates, suggesting the presence of races of this pathogen in lentil (Barilli and Rubiales 2023).

Little is known on variation on virulence among remaining rust species. Consequently, further studies on host range and particularly on virulence of most legume rusts are needed. Understanding the host range and diversity of rust isolates is crucial to design efficient management strategies and to identify alternative hosts that can serve as reservoirs to propagate the disease and cause unexpected outbreaks. Moreover, if the same rust species or pathotype can infect two legume species, knowledge about the resistance developed for one of these species can also be useful for improving resistance in the other (Kawashima et al. 2016).

Disease management

The very efficient spreading mechanism of rust that allow the transport of its urediniospores by winds or travellers over thousands of kilometres, coupled with its wide host range, make the eradication of the pathogen in the field a challenging task. Efficient control of rust requires the integration of different disease management approaches. The elimination or reduction of the pathogen propagules and of its aerial dispersion are the primary objectives of these disease control measures (Chandrashekara et al. 2022).

Cultural control

Agricultural practices can play a major role in reducing rust incidence in legumes. Accordingly, they are included in a sustainable integrated pest management approach in combination with other strategies including biological and chemical control methods. Field assessment of previous crop, tillage, sowing date, crop** system, plant density and weed control are included as parameters that potentially decrease rust severity (Juroszek and von Tiedemann 2011).

Early sowing can facilitate the premature dispersal and germination of fungal spores, prompted by the advent of favourable temperatures and humidity levels. This phenomenon triggers an early emergence of the disease and increase the number of fungal cycles during the epidemic phase, thereby intensifying crop damages (Dawit and Andnew 2005; Das et al. 1999). Accordingly, postponing planting was shown to diminishes the impact of rust in peanut (Das et al. 1999), faba bean or pea (Eshetu et al. 2018; Singh et al. 2014). It was also efficient in cereals against some Uromyces species (Dawit and Andnew 2005), making it a suitable solution to prevent or reduce rust incidence in the field. The preceding crop in rotation may also plays a pivotal role in the severity of rust infestation on the subsequent legume crop. If this crop is susceptible to the same rust species, the remaining debris might serve as an inoculum reservoir, initiating plant infection under conducive environmental conditions.

The agricultural system, such as monoculture or mixed crop**, is another determinant of rust severity under field conditions. Recent studies underscore the superiority of intercrop** over monoculture in managing diseases in legumes (Singh et al. 2014; Guo et al. 2021; Luo et al. 2022; Zhang et al. 2019). In intercrop** systems, legumes are mixed with other crops such as cereals. These grasses, generally taller than legumes, can form a barrier which impedes spore dispersion, thereby curtailing the number of fungal reproductive cycles in the epidemic stage and subsequently diminishing the disease impact on host plants (Guo et al. 2021; Villegas-Fernández et al. 2023). Moreover, crop** mixtures can efficiently reduce weeds in the field that often serve as alternative hosts for rust, which also contribute to lessens disease severity in the field (Shtaya et al. 2021). Controlling the alternative host is also a key strategy in managing rust in legumes, as evidenced by the historical success in controlling wheat stem rust by eradicating barberries (Zhao et al. 2016). This method reduces the likelihood of sexual recombination of the rust pathogen, which often occurs on its alternate host. For instance, U. pisi, the pea rust pathogen, completes its life cycle on Euphorbia cyparissias and E. esula, which can grow in the vicinity of pea fields as spontaneous weeds and spread the fungal aeciospores over the crop (Pfunder and Roy 2000). By managing or eradicating such alternate hosts, the source of inoculum is significantly reduced, thereby reducing the spread of the disease. This approach is crucial in integrated pest management programs.

Lastly, the density of plantations may influence the spread and severity of the fungus (Fernández-Aparicio et al. 2006). While certain studies observed a strong correlation between sowing density and rust severity (Eshetu et al. 2018; Fernández-Aparicio et al. 2006; McEwen and Yeoman 1989), others only detected a marginal effect, if any, with climatic conditions playing a more substantial role in the disease development (Olle and Sooväli 2020; More et al. 2020).

Biological control

Biological control has primarily revolved around the use of living organisms such as predators, parasites, and pathogens to manage pest and disease populations in the field. However, the scope of biological control also encompasses the utilization of natural compounds derived from these organisms. This includes pheromones, hormones, and metabolism-derived products that may function as repellents, attractants, or growth inhibitors. Both approaches have been documented in legume crops to manage rust disease.

Various microorganisms have been identified as biological control agents (BCAs) against rust disease. The study of host plant leaf microbiome allowed the discovery of antagonistic endophytic bacteria (Kiani et al. 2021). These beneficial bacteria limit fungal growth thereby reducing rust disease severity (Yuen et al. 2001). Several studies have demonstrated the effectiveness of bacteria from the Bacillus genus, against both cereal and legume rusts where they significantly mitigate disease severity in the field and under controlled conditions (Baker 1983; Baker et al. 1985). In addition, bacteria from the Pseudomonas genus have also been described as potential BCAs against U. appendiculatus (Abo-Elyousr et al. 2021). Endophytic fungi able to induce resistance, antagonize rusts or even act as hyperparasites of the pathogen have also been described (Fontana et al. 2021). For instance, different strains of Trichoderma spp. have been reported to stimulate systemic resistance in common bean against U. appendiculatus (Burmeister and Hau 2009; Cruz-Triana et al. 2018). These strains can also antagonize the pathogen by inhibiting urediniospore germination and germ tube growth (Abeysinghe 2009). Further studies highlight the use of BCAs from Simplicillium and Cladosporium genera as hyperparasites of various rust species from the genera Puccinia, Phakopsora, and Uromyces genera (Si et al. 2022; Assante et al. 2004; Moricca et al. 2005; Barge et al. 2022).

A more accurate approach of biological control involves unveiling which specific products from these fungal and bacterial species can limit the growth and spread of rusts. This method would restrict the introduction of exogenous organisms into crop fields, potentially averting ecosystem destabilization if their growth becomes uncontrolled (Herskowitz et al. 2023). In this direction, the inhibitory effect of several compounds isolated from plant essential oils or fungal secondary metabolism have been validated against different rust species. The antifungal effect of crude plant extracts has also been tested. For instance, crude extracts from Ageratum conyzoides reduced about 50% the number of P. arachidis pustule in peanuts (Yusnawan and Inayati 2018). Likewise, a 50% reduction in the germination of U. viciae-fabae urediniospores was observed following the application of crude extracts from Pelargonium zonale (El-Fawy et al. 2021) while Nigella sativa extract reduced common bean rust severity by up to 96% (Arslan et al. 2009). Evidence also indicates that essential oils derived from plants can decrease the number of rust pustules in legumes. Application of linseed oil, for example, was found to completely inhibit U. appendiculatus urediniospore germination in both in vivo and in planta experiments (Arslan 2014). Other trials on U. viciae-fabae demonstrated a reduction of rust severity of up to a 96% after treating infected plants with basil oil three hours after inoculation (Oxenham et al. 2005) while hyssop and pumpkin seed oils were less efficient (Letessier et al. 2001; El-Fawy et al. 2022). While not achieving as significant a reduction in rust severity on the entire plant, the application of these essential oils offers additional synergistic benefits for legumes, potentially enhancing plant height and yield, as demonstrated under greenhouse conditions (El-Fawy et al. 2022).

The isolation and analysis of secondary metabolites have also constituted an important area of research in the control of rust. Effective substances produced by the secondary metabolism of specific bacteria have been identified, with notable contributions from the Bacillus (Lim et al. 2017; Manjula et al. 2004), Trichoderma and Cladosporium genera (El-Hasan et al. 2022; Nasini et al. 2004). Similarly, several bio-compounds isolated from the secondary metabolism of phytopathogenic fungi such as Seiridium cupressi, Diplodia quercivora, and Ascochyta lentil was shown to reduce the U. pisi severity on peas (Barilli et al. 2016, 2017, 2022). Additionally, evidence suggests that the accumulation of phytoalexins in non-host species inhibits the development of the pathogen within their plant tissue. For instance, phytoalexins such as medicarpin and scopoletin have been associated with resistance to P. pachyrhizi in the non-host species M. truncatula and A. thaliana, respectively, and their influence on soybean plants have been successfully tested against P. pachyrhizi (Beyer et al. 2019; Ishiga et al. 2015). Accumulation of these phytoalexins can also be induced by exogenous compounds that induce the systemic acquired response (SAR), such as the elicitors BTH (Benzo[1,2,3]thiadiazole-7-carbothionic acid) and BABA (D, L-β-aminobutyric acid), as it has been demonstrated in pea against U. pisi (Barilli et al. 2010a, 2012a, b, 2015). This line of research highlights a wide range of biological resources to enhance plant defence mechanisms against rust pathogens in an eco-friendlier way although their large-scale application in the field is so far not possible being compromised by the low yield of the isolation methods and the lack of complete study of their environmental impact.

Chemical control

The management of rusts in agricultural fields can be effectively achieved using chemical-synthesized fungicides. Among the primary phytochemicals employed for rust disease control, albeit with varying efficiencies, are triazoles, strobilurins, and carboxamides (Juliatti et al. 2017; Chen 2005; Alam et al. 2007). Triazoles function by inhibiting the enzyme 14α-sterol demethylase, thereby obstructing the binding of ergosterol, and subsequently disrupting the structural and functional integrity of the fungal cell wall. Additionally, triazoles exhibit systemic action, disseminating through the leaves (translaminar movement) and the xylem (acropetal movement). Their effectiveness, whether applied alone or in combination with benzimidazoles, has been validated in several legume crops against rust species such as U. viciae-fabae, U. appenditucalus, and U. lupinicolus (Devi et al. 2020; Emeran et al. 2011; Etheridge and Bateman 1999; Modesto et al. 2005; Sugha et al. 2008). However, natural mutations in sterol demethylase have been identified in some cereal pathogen which reduced their sensitivity to triazole. Therefore there is intermediate risk that the legume rust pathogens develop triazole resistance (Cools et al. 2006).

Strobilurins impede mitochondrial respiration by targeting the electron transfer chain between cytochromes b and c1, which hampers ATP synthesis (Köhle et al. 1997). Known as QoIs (quinoline outside inhibitors), these broad-spectrum fungicides have demonstrated efficacy against various rust species and, in some instances, have enhanced growth and yield (Rasha and Mohamed 2021; Glaab and Kaiser 1999). Although the risk of rust species develo** resistance to strobilurins is generally low, studies have indicated that a single point mutation in the mitochondrial cytochrome b (the target of strobilurins) can lead to QoIs fungicide resistance (Grasso et al. 2006a, b).

The action mechanism of carboxamides targets the enzyme succinate dehydrogenase (SDH), a crucial component of the tricarboxylic acid cycle (TCA) and the mitochondrial electron transport chain (Rheinheimer 2019). Carboxamides, thus, inhibit fungal cell respiration by blocking the TCA cycle at the oxidation stage from succinate to fumarate, culminating in the rapid cell death. These fungicides have also been tested against rust in legumes; for instance, they limit rust damage in faba beans affected by U. viciae-fabae (Emeran et al. 2011). While carboxamides effectively mitigate rust disease across a spectrum of legumes and cereal crops, the Fungicide Resistance Action Committee (FRAC) has developed resistance management recommendations for various crop pathogens to minimize the risk of resistance to this class of fungicides (Sierotzki and Scalliet 2013).

Despite their efficiency, the utilization of fungicides imposes a significant financial burden on legume production, especially in develo** countries where legumes are the main protein source for human food (Emeran et al. 2011). The use of these chemicals can also pose health risks to users, adversely impact the environment, and lead to the emergence of fungicide-resistant rust strains (Oliver 2014). Consequently, cultivating varieties with an adequate level of durable resistance represents the most effective strategy for rust disease control in legumes.

Basis of rust resistance

Rust infection cycle

The study of the resistance mechanisms developed by the host plant to respond to the pathogen are intricately linked to the infection process of rust. This infection cycle, detailed in the Fig. 2, begins when a spore lands on the tissue surface (Fig. 2a). Upon recognition of the leaf surface, the rust spore adhere to the cuticular surface stabilized by hydrophobic interactions (Hahn 2000). If the environmental conditions of temperature and humidity are favourable to reactivate the spore metabolism (More et al. 2018), it will form a germ tube which move through the leaf employing a touch-responsive process known as thigmotropism. If it successfully locates a stoma, cell morphogenesis occurs in the apical tip of the germ tube where the cytoskeleton and microtubules reorganized to form an appressorium over it, serving to penetrate the host (Fig. 2b). When the appressorium is fully differentiated and delineated from the cytoplasmatic-empty germ tube by a septum, the first round of divisions of the two nuclei has been completed (Hahn 2000). The appressorium are adhered to the stomata thanks to cell wall proteins called hydrophobins which form amphipathic layers in fungal cell walls to ensure an effective penetration. Successful penetration of the stomata leads to the formation of a substomatal vesicle in the substomatal space (Fig. 2c). From here, an initial infection hypha emerges and, upon contacting a mesophyll cell, can differentiate at its tip into a haustorial mother cell (HMC). Then the pathogen can enter the mesophyll cell via a neckband and forms a haustorium which is the feeding structure of the pathogen (Fig. 2d). If the host plant does not activate any hypersensitive response, this first haustorium is considered effective and starts hijacking nutrients from the host cells. Then, the fungus will continue to develop secondary hyphae to infect more cells and continue extracting nutrients to expand the colony within the host (Fig. 2e). Once it has accumulated sufficient resources, sporogenic tissue begins to form, and spores emerge through an opening on the tissue surface created by pressure exerted from within (Fig. 2f), leading to what we know as a pustule. This detailed understanding of the rust infection process is crucial for develo** effective resistance strategies in the host plants, as it reveals the critical stages where intervention might be most effective.

Fig. 2
figure 2

Rust infection process: a spore deposition; b spore germination and appressorium formation; c stomatal penetration and substomatal formation; d development of the first intracellular hyphae and haustorium formation: e colonization; f spore formation and release. Created with https://www.BioRender.com

Full size image

Resistance mechanisms against rust

The detailed monitorisation of the infection process at microscopic level on a set of differential accessions allowed the detection of different resistance mechanisms. The first resistance mechanism expressed by the plant is related with leaf morphology, reducing surface area and growth orientation, thus hindering spore deposition. It has been proposed that spore deposition (Fig. 2a) would be lower on plants with vertical leaves compared to horizontal ones. However, the impact of leaf orientation on spore deposition is less significant than the effect of the foliar microclimate on germination. It has also been observed that bean rust germ tubes grow along trichomes on soybean plants, reducing the contact with the leaf surface (Wynn 1975). In this case, the trichome acted as a passive screen that can reduce disease severity although, this factor was not decisive in final plant severity under field conditions (Mmbaga et al. 1994). Some additional pre-infection mechanisms have been described including those preventing pathogen adhesion to the leaf surface (Mmbaga et al. 1994; Wynn and Staples 1981), diverting the thigmotropically sensitive germ tube away from stomata (Wynn and Staples 1981), impeding stomata recognition through atypical morphology of stomatal guard cells (Wynn 1975), and the secretion of protective compounds on the leaf surface (Niks and Rubiales 2002; Prats et al. 2007). Genotypic differences that limit spore germination and directional growth of the germ tube have been detected against U. pisi and U. viciae-fabae (Vaz Patto et al. 2009). However, these mechanisms can, at best, reduce infection levels at the early epidemic cycles but are of marginal importance in natural conditions (Rubiales and Moral 2004; Vaz Patto et al. 2009).

For pathogenic fungi that penetrate stomata, it is important that they locate them through an efficient chemical recognition (Cooper et al. 2007). The U. viciae-fabae fungus seems quite inefficient since typically only about 50% of germ tubes find a stoma in faba beans, compared to figures ranging from 80 to 100% for pea rusts caused by U. pisi (Sillero and Rubiales 2002; Barilli et al. 2009c). Germ tubes of rust that reach a stoma must stop growing and develop an appressorium to enter the leaf (Fig. 2b). Reaching a stoma does not automatically result in stoma recognition and subsequent penetration (Chethana et al. 2021). There is evidence that urediniospore germs of U. appendiculatus detect morphological features of host stomata on which to develop appressoria in common bean (Wynn 1975; Hoch et al. 1987). Generally, stoma recognition is very efficient: typically, more than 90% of germ tubes reaching a stoma form an appressorium. However, there are evidence of misplaced appressorium formation by the faba bean rust fungus, U. viciae-fabae. In this pathogen, the proportion of misplaced appressoria reach about 20%, but it is rather uniformly distributed among faba bean accessions offering little opportunity for breeding (Sillero and Rubiales 2002).

Despite the significance of pre-penetration mechanisms in some genotypes, a study in pea showed that their expressions have little impact on rust epidemiology in the field (Barilli et al. 2009c). The most efficient resistance mechanisms against rust occur after the formation of the substomatal vesicles. Non-host resistance, an inherent defence mechanism against non-adapted pathogens, is typically manifested before the formation of the first haustorium (Bettgenhaeuser et al. 2014). This type of resistance is also relevant to host-pathogen interactions, significantly contributing to partial resistance (PR), which is a quantitative mechanism characterized by a slower rate of disease development compared to susceptible plants (Parlevliet 1979). Being controlled by multiple genes, this mechanism showed an enhanced durability compared to monogenic R gene-mediated resistance (Niks and Rubiales 2002; Rubiales and Niks 1995). PR has been documented in a range of legumes including faba bean (Sillero and Rubiales 2002) and M. truncatula (Rubiales and Moral 2004), where a considerable proportion of infection units fail to establish haustoria. By contrast, host resistance mediated by R genes is characteristically induced after haustoria formation and is frequently correlated with the hypersensitive response, a localized cellular apoptosis designed to constrain pathogen propagation (Camagna and Takemoto 2018). Despite the efficacy of this response, the durability of R gene-mediated resistance is potentially compromised due to the pathogen evolutionary capacity to overcome specific resistance genes (Niks and Rubiales 2002).

The spectrum of rust resistance responses in legume crops are mostly categorized as incomplete. Although the setup of the different resistance mechanisms reduce and delay rust pustule emergence, they are not able to completely stop the pathogen life cycle. For instance, some faba bean and pea cultivars showed an incomplete level of partial hypersensitive resistance (HR) permitting some degree of sporulation in the presence of host cell necrosis surrounding the infection site (Sillero et al. 2000; Osuna-Caballero et al. 2022).

Non-hypersensitive resistance constitutes an alternative form of incomplete resistance that impedes epidemic progression without eliciting programmed cell death. Research indicates that this type of rate-reducing resistance is widespread within legumes species (Sillero et al. 2012; Barilli et al. 2009a; Osuna-Caballero et al. 2022; Singh et al. 2015). This situation contrasts with other pathosystems including cereal rusts that are dominated by complete HR-based resistance and for which such PR have been rarely described (Niks and Rubiales 2002).

Genetic basis of resistance against rust

Precision genetic breeding for rust resistance requires an understanding of the genetic basis of resistance. The qualitative and quantitative resistance mechanisms detected so far in legumes are dependent on one or multiple genes, respectively. This distinction between single gene and multiple resistance genes is crucial, as it influences the breeding methodology and the potential durability of the resistance. Linkage map** and GWAS have enabled the identification of specific genetic loci and alleles responsible for resistance, thereby providing a foundation for targeted breeding strategies.

Monogenic resistance in rusts enable the cell programmed death when the haustorium forms inside the host. Although it is not the most common source of resistance in legumes, its use in breeding is available for some species. The known candidate resistance genes or QTLs and their location are displayed in Table 2 when the resistance sources contribute > 10% to the phenotypic variance.

Table 2 Quantitative trait loci (QTLs) and candidate resistance genes against rust infecting legumes, their genetic location (linkage group or chromosome) and linked markers
Full size table

In pea, only incomplete resistance has been identified against U. viciae-fabae and U. pisi. While some QTLs associated with resistance have been mapped, they are not yet suitable for marker-assisted selection (MAS) in breeding programs (Rubiales et al. 2011). There is evidence suggesting that partial resistance to U. viciae-fabae may be due to a single major gene (Ruf), with two RAPD markers identified nearby, but not close enough for effective MAS (Vijayalakshmi et al. 2005; Rai et al. 2011, 2016). For U. pisi, a QTL responsible for 63% of resistance was located, with two associated RAPD markers, but further validation in different environments and genetic backgrounds is needed before these findings can be applied to MAS (Barilli et al. 2010b, 2018). In those cases, the development of standard markers and conversion of RAPDs to sequence-characterized amplified regions (SCARs) is necessary to improve their utility for MAS. On the other hand, new silico DArTseq makers have been associated to U. pisi partial resistance in pea where putative genes were proposed as resistance’s causative agents for their use in breeding programs, but validation is still needed (Osuna-Caballero et al. 2024).

Likewise, researchers have found only incomplete resistance to rust caused by U. ciceris-arietini in chickpea (Sillero et al. 2012). A major QTL, accounting for 81% of the resistance in adult plants, was mapped to linkage group 7 on the chickpea genetic map in an interspecific cross population (Madrid et al. 2008). Interestingly, this resistance, even when monogenic should be regarded as incomplete non-hypersensitive resistance, caused by pre-haustorial resistance mechanisms, resulting in incomplete sporulation retarding disease progress. This resistance is thought to be controlled by a single gene (Uca1/uca1) closely flanked by two STMS markers suitable for reliable marker-assisted selection for rust resistance in chickpea breeding programs (Madrid et al. 2008).

In lentils, PR resistance to U. viciae-fabae is very common, but HR has also been described (Rubiales et al. 2013b; Negussie et al. 2005, 2012; Barilli et al. 2023). Monogenic resistance has been reported and research is advancing on identifying its chromosomal location and linked markers (Kant et al. 2004; Dikshit et al. 2016; Fikru et al. 2014; Saha et al. 2010; Singh et al. 2021). However, this HR resistance in lentils could not be associated to race-specific U. viciae-fabae isolates since these races have not been described so far (Barilli et al. 2023). In contrast, Significant association between the partial resistance and a specific SRAP and SSR markers has been found and could be used in MAS, though identification of markers closer to the gene would improve this approach (Kant et al. 2004; Dikshit et al. 2016; Fikru et al. 2014; Saha et al. 2010; Singh et al. 2021).

In common bean, many sources of HR have been identified that are specific to individual races of U. appendiculatus (Hurtado-Gonzales et al. 2017b). As detailed in Table 2, the QTLs Ur9, Ur5, Ur-Dorado, Ur-Ouro Negro, Ur14, Ur3, Ur6, Ur7, Ur11, Ur13, Ur4, Ur12 confer resistance to the U. appendiculatus races with the same name. However, some of these resistance sources have been overcome by the pathogen adaptation (Hurtado-Gonzales et al. 2017a). Recent studies have advanced our understanding of the genotypic basis of common bean resistance to diverse rust strains (Wu et al. 2021b). Hypersensitive resistance in faba bean, controlled by major-effect genes, has also been identified (Avila et al. 2003). For instance, three RAPD markers linked to a the gene Uvf-1 confer hypersensitive resistance to the U. viciae-fabae race 1 which is also associated with two additional markers identified in repulsion phase (Miklas et al. 1993).

In cowpea, four QTLs have been proposed to induce resistance against rust with twelve linked markers available for MAS (Li et al. 2007; Wu et al. 2018a). Recently, the first QTL proposed, Rr1 was coincident with the Ruv1 in the Wu et al. (2018a) studies. Therefore, the SNPs markers are considered the most reliable in their use for breeding (Wu et al. 2018b).

In the case of pigeon pea, a single source of partial resistance against P. pachyrhizi known as the CcRpp1 gene (Cajanus cajan Resistance against Phakopsora pachyrhizi 1) has been reported and successfully transferred to transgenic soybean plants (Kawashima et al. 2016).

Gene regulation upon rust infection in legumes

A transcriptional profiling during plant–pathogen interaction allows identifying candidate resistance genes from the host plant and genes involved in infection processes from the pathogen (Jha et al. 2021). Rust triggers important transcriptional changes in legume plants during infection, with several hundreds of genes being either up- or downregulated. RNA-seq and transcriptome analyses are powerful tools to identify key defence responsive genes and transcription factors in legume-rust interaction.

Defence/resistance genes

In peas, an increase in glucanase activity has been observed during infection by U. viciae-fabae and U. pisi (Barilli et al. 2010a; Kushwaha et al. 2018). This enzyme activity is thought to contribute to the formation of phenolic compounds involved in lignin formation in cell walls, thereby strengthening the plant defence against the pathogen (Yadav and Chattopadhyay 2023). This aligns with transcriptomic analyses in grass pea, where resistance to U. pisi was correlated with the overexpression of an endo-beta-1,3-glucanase gene in resistant genotypes. Additionally, overexpressed genes in resistant genotypes of grass pea suggested a comprehensive molecular response to rust infection which has also been indicated in partially resistant pea accessions (Almeida et al. 2014; Santos et al. 2018). These previous studies in grass pea indicated the overexpression of genes related to phytohormones and transcription factors in resistant genotypes, suggesting a shared genetic basis for resistance in these related legume species against the same rust pathogen and thereby enabling more breeding opportunities (Martins et al. 2022; Osuna-Caballero et al. 2024). In the case of V. angularis, a relative to V. unguiculata, a response to U. vignae infection was characterized by the activation of genes encoding glutamate receptor proteins (Yin et al. 2023).

Additionally, genetic expression studies in common bean inoculated with U. appendiculatus race 53 revealed significant changes in over five hundred genes when compared to control conditions at various time points post-inoculation (0, 12, and 84 h) (Ayyappan et al. 2015). Among these, 90 genes were differentially expressed at all time points, including genes involved in stress responses such as calmodulin, cytochrome P450, chitinase, DNA polymerase II, and LRR, as well as transcription factors including WRKY, bZIP, MYB, HSFB3, GRAS, NAC, and NMRA (Ayyappan et al. 2015). These findings underscore the importance of these genes in the common bean-rust interaction which are also similar to those found in other rust species, such as peanut rust caused by P. arachidis (Ayyappan et al. 2015; Rathod et al. 2020). Similar differential expression patterns were observed in the “Sierra” cultivar of common bean harbouring the resistance gene Ur3, where genes containing NBS, LRR, and TIR signature motifs, along with WRKY-type transcription factors, were overexpressed at the onset of infection (Todd et al. 2017). Differential expression was also assessed in common bean to study the genetic architecture of Ur4 resistance source. In that study, up to 90 genes were differentially expressed during U. appendiculatus infection. They were mainly attributed to stress response, hormone regulation/signalling, transport, and cell wall formation (Thibivilliers et al. 2009). In faba bean, a study demonstrated a systemic plant response to localized leaf infection by U. viciae-fabae, involving changes in carbohydrate and amino acid metabolism as an adaptive strategy to the pathogen’s entry into cells via the haustorium (Wirsel et al. 2001) which agrees with similar studies in U. appendiculatus (Puthoff et al. 2008).

Overall, several candidate resistance genes have been identified in plants by transcriptomic approaches which can complement the previously described genetic insights for the breeding of rust-resistant crops. However, functional studies are still missing to validate their role during plant–pathogen interaction.

Understanding both sides of the plant–pathogen interaction is important to completely unravel the molecular basis of resistance. However, most previous transcriptomic studies targeting legume–rust interaction, only compared the plant transcripts induced in response to the pathogen due to low detection of fungal transcript. Identification of rust effector genes controlling pathogen host colonization must be considered in transcriptomic studies focused on the identification of differentially expressed genes from the rust side and integrated with comparative rust genomics studies.

Candidate effector molecules

A significant focus in the search of effector proteins of legume rust pathogen, has been placed on U. appendiculatus, P. pachyrhizi, and U. viciae-fabae which genomes are available (Link et al. 2014a). In particular, The broad bean-U. viciae-fabae and common bean-U. appendiculatus pathosystems have been extensively studies as model systems for in-depth investigation of legume-rust interactions and asian soybean rust, respectively (Link and Voegele 2008; Thibivilliers et al. 2007). Preliminary studies on U. appendiculatus have analyzed the haustoria transcriptome to predict and identify effector molecules secreted by the pathogen (Link et al. 2014b). Among the discovered effectors of U. appendiculatus, some effectors including Uaca_9, Uaca_12, Uaca_14, and Uaca_22, can suppress the hypersensitive response in the host plants (Qi et al. 2019). Additionally, effectors like Uaca_4, Uaca_5, Uaca_7, Uaca_9, Uaca_28, and Uaca_44 can suppress basal resistance in Nicotiana benthamiana against the bacterial pathogen Pseudomonas syringae (Qi et al. 2019). This indicates the broad impact these effectors might have across different host and pathogen species. It was also noted that some of the genes encoding these effector proteins contain highly conserved motifs within the Pucciniales family, suggesting that the effector secreted by the different rust species and their infection process are likely similar (Cooper et al. 2016). Accordingly, the infection process of both U. appendiculatus and P. pachyrhizi by the secretion of families of hydrolase proteins to degrade the host plant’s cell wall (Cooper et al. 2016). They also produce structural proteins crucial for forming and stabilizing the haustorium within the host cells, which is a pivotal step in the infection cycle (Link et al. 2014b). The effector proteins from these pathogens are localized in the cytoplasm and nucleus of the host plant cells, where they exert their influence by activating or suppressing various plant responses (de Carvalho et al. 2017).

In the case of U. viciae-fabae, specific molecules responsible for the biotrophic interaction between the pathogen and its host have been identified (Voegele 2006). One such molecule, Uf-RTP1p (Rust Transferred Protein 1), secreted by the haustorium, is found within the infected cells, including their nucleus (Kemen et al. 2005). Molecules with homologous domains have been discovered in other rust species and proposed to have analogous functions (Kemen et al. 2005; Vieira et al. 2012). The detection of these molecules facilitates accurate quantification of haustoria using RT-qPCR analysis (Voegele and Schmid 2011). Therefore, a deeper understanding of the substances secreted by the fungus and their functional roles enhances our knowledge of the effectors and the biotrophic interaction between rust fungi and its host (Link et al. 2005; Voegele et al. 2005). This knowledge is crucial for develo** specific improvement strategies to breed fully resistant plant to rust since the most effector genes are known; new resistance genes could be identified (Jakupović et al. 2006; Link 2020).

Breeding for resistance

The main objective of breeding for rust resistance is to develop varieties that either show delayed disease onset, minimal symptom development, or slow disease progression, thereby minimizing crop damage. This process begins with the crucial step of identifying and characterizing potential resistance sources for integration into breeding programs.

Resistance screening methods

Efficient screening methods are essential for discovering new resistance sources against rust. In general, this process starts with mass screenings, where large germplasm collections, primarily from the same legume species or occasionally from its wild relatives are evaluated. These initial screenings aim to identify potential resistance sources. Following this, the resistance mechanisms of these promising candidates are further investigated through more in-depth screenings on a selected group of accessions. This two-tiered approach allows breeders not only to identify novel resistance sources but also to understand the underlying mechanisms, thereby aiding the development of rust-resistant varieties.

Mass screenings can be conducted either in natural field settings or under controlled conditions. It is a cornerstone in identifying aerial disease resistance in legumes (Sillero et al. 2006). These screenings employ a range of tools and techniques, to monitor symptom development on the whole plant or on the leaves, which are the most affected plant organs. Field screenings enable the simultaneous evaluation of extensive germplasm collections in conditions where natural inoculum is present allowing an understanding of the genetic and environmental factors that influence the phenotypic variances (Civantos-Gómez et al. 2022; Das et al. 2019). For more uniform and precise assessments, artificial inoculation with urediniospore is often employed and recommended when natural infestation is not high enough. This approach consists of spraying the plants with aqueous suspension of rust spores or dusting mixture of spores in an inert carrier, ensuring consistency across the experimental trial (Sillero et al. 2000). However, natural conditions present challenges, such as the co-occurrence of other aerial diseases like ascochyta blight or powdery mildew, which can complicate rust assessments and lead to the underestimation of its impact providing some risks of confusing escape with resistance (Porta-Puglia et al. 1993). To mitigate this, repeating inoculation with urediniospore may be necessary during field trials to ensure accurate evaluation of legume responses to rust (Barilli et al. 2009a). In addition, it is recommended to inoculate the plants after sunset to benefit from both the darkness and the high relative humidity of the night (Sillero et al. 2006).

Under field conditions, the assessment of quantitative and race-specific resistance to rust involves several methodologies. A common approach includes the visual estimation of foliar area affected by pustules, referred to as disease severity percentage (DS). When periodic evaluations of DS are performed, it provides insights into both the final severity of the disease (captured in the last DS data) and its progression overtime. Rust disease progression is estimate through the integration of the periodic DS evaluation into some parameters including the area under the disease progress curve (AUDPC) and the epidemic growth rate (r) (Arneson 2001; Jeger and Viljanen-Rollinson 2001), both critical in understanding the rust impact over time. Alongside DS, it is standard to record the infection type (IT) in the field. Various measurement scales have been developed for IT, which describe the plant reaction to rust disease. This reaction is characterized by the extent of necrosis or chlorosis at the infection sites, as well as the sporulation rate of the colonies that have formed on the tissue. One of the most widely used IT scales, developed for wheat rusts by Stakman et al. (1962), categorizes the reaction as follows: 0 = no symptoms,; = necrotic flecks, 1 = tiny pustules without sporulation, 2 = necrotic halo surrounding small pustules, 3 = chlorotic halo surrounding pustules, and 4 = well-formed pustules without associated chlorosis or necrosis. On this scale, values between 0 and 2 indicate resistance, while 3 and 4 imply susceptibility. More comprehensive scales such as the scale developed by Bernier et al. (1984) for faba beans rust, combine IT with DS percentages for refined field assessments. This scale, for instance, ranges from 1 (highly resistant) to 9 (highly susceptible), considering both the leaf area and the whole plant affected. Similar scales exist for another legume rusts such as peanut and pea (Sokhi et al. 1984; Subrahmanyam et al. 1995). In many pathosystems traditional visual assessments are increasingly being supplemented, and in some cases replaced, by remote sensing technologies. These methods, leveraging the contrast between damaged and healthy tissue, are develo** new models for rust evaluation with comparable or higher accuracies than visual estimates (Simko et al. 2017). Most current models have been targeting cereal rusts, where resources and research investment are more substantial (Rubiales et al. 2023). However, adapting these remote sensing methodologies for rusts in legumes could enhance precision and allow timesaving in field evaluations under natural conditions.

A more detailed analysis of disease symptoms is feasible under controlled conditions using growth chambers or greenhouses. This setting enables evaluations at both seedling and adult plant stages. These systems are essential to test the efficacy and multiplication of rust isolates for subsequent evaluation. The most common method for inoculation involves dusting plants with urediniospores diluted in an inert carrier such as talc powder (Chand et al. 2004). To achieve uniform spore deposition, the use of inoculation towers is recommended (Sillero et al. 2000). After inoculation, it is necessary to provide the environmental conditions required for urediniospore germination and successful plant infection. This typically involves incubation in darkness with 100% relative humidity for 12–24 h (More et al. 2018). Depending on the rust species, the first macroscopic symptoms are visible between 7 and 9 days post-inoculation (dpi), marking the start of symptom evaluations. Unlike field evaluations, where DS is the most recorded parameter, controlled conditions allow for a more in-depth analysis of the disease and the measurement of additional parameters. This includes counting pustules per unit area (infection frequency, IF), typically in a defined leaf area, and measuring pustule size in mm2 (Asare et al. 2019; Barilli and Rubiales 2023). IF also facilitates the calculation of the latency period (LP) that is the time between inoculation and observation of 50% of total pustules. Both DS percentage estimation and pustule counting is highly time consuming, especially when assessing large plant collections (Bock et al. 2020). Therefore, standard area diagrams (SADs) have been developed for some rust pathosystems to help reduce evaluator bias (Del Ponte et al. 2017). However, these SADs have not been adapted for most legume-infecting rust species, except for soybean rust (Franceschi et al. 2020). Efforts have also been made to automate the evaluation process under controlled conditions. For instance, through detached leaf assays, IF and DS can be calculated using easily acquired images such as RGB (red-green-blue) or thermal sensors, as shown for pea and soybean rusts (Alves et al. 2022; Olivoto 2022; Osuna-Caballero et al. 2023). These develo** techniques could be readily adapted to other rust pathosystems, as they share similar symptoms, improving evaluation precision and enabling the assessment of large germplasm collections.

In addition, the common infection cycle shared by most rust species allow the evaluation of the different stages of the infection process by microscopic observation of infected leaves at an early stage of the interaction. For instance, in several rust-legume pathosystems, the formation of appressoria over the stomata has been assessed, allowing for the calculation of tissue penetration percentage (first 3–6 h post-inoculation) (Fig. 2b). Substomatal vesicle formation, hyphal development, and haustorium formation within cells (6–12 h post-inoculation) have also been evaluated in some cases (Fig. 2c) (Dugyala et al. 2015; Sillero and Rubiales 2002). Colony size within plant tissue at 24- or 48-hours post-inoculation and the extent of cell death surrounding the rust colony are others commonly assessed parameter (Fig. 2d) (Barilli et al. 2009c; Kushwaha et al. 2016; Negussie et al. 2012; Sillero et al. 2012). These microscopic evaluations allow to characterise the resistance mechanisms expressed by the most resistant accessions and select cultivar with a specific resistance mechanism for breeding.

Conventional breeding

Classical breeding techniques such as backcrossing, pedigree selection, and recurrent selection can be used to develop rust-resistant cultivars for the efficient management of rust in the field. These techniques involve crossing elite cultivars with sources of rust resistance. The development of rust resistant legume cultivars with interesting agronomic potentials is a main goal of any breeding programs (Renzi et al. 2022). The search and identification of resistant sources is the first step in any classical breeding program for rust resistance.

Once the source of resistance is identified, the introgression of the genomic regions conferring rust resistance into non-resistant elite genotypes can be performed through a complex crossing selection scheme. For instance, Conventional breeding for rust resistance in peanuts involves introgression of resistance genes from wild species into cultivated varieties (Stalker 2017). This has been achieved through wide hybridization, where genes from cross-compatible wild species are transferred into the cultivated peanut. Efforts in India led to the development of rust-resistant peanut breeding lines such as VG 9514, derived from A. cardenasii which were subsequently used as parental lines to develop map** populations for genetic and QTL map**, enabling the development of molecular markers for rust resistance selection (Varman 1999).

Rust resistance breeding in soybean has involved conventional strategies such as gene pyramiding (Chander et al. 2019). This involves combining multiple Rpp genes within a single genotype for broader and more durable resistance (Yamanaka et al. 2015). Rpp gene pyramiding have been facilitated by the use of molecular markers that contributed to identify and select these genes and monitor their transfer. Studies have shown that combinations of Rpp genes, like Rpp2, Rpp3, Rpp4, and Rpp5, provide enhanced resistance to P. pachyrhizi (Meira et al. 2022; Yamanaka and Hossain 2019). However, the effectiveness of these gene combinations varies with the genetic background of the soybean variety, indicating the importance of considering both specific genes and the overall genetic context in breeding programs.

In common bean, the optimal traditional breeding approach also involves gene pyramiding to accumulate various race-specific resistance sources (Beaver et al. 2003), considering the isolate’s specific climatic zones. Common bean varieties with rust resistance are typically developed through crossbreeding, backcrossing, and continuous disease pressure over successive generations, ensuring the acquisition of homozygous resistance genes. Breeding efforts have resulted in the release of resistant cultivars to different races of U. appendiculatus carrying different resistant genes, alone or in combinations (Beaver et al. 1999, 2015, 2020; Osorno et al. 2021; Pastor-Corrales et al. 2007).

In the case of chickpea, a single interspecific cross between C. arietinum and C. reticulatum enabled the wild parent to contribute the Uca1/uca1 gene, resulting in segregating resistance lines within the offspring (Madrid et al. 2008). This allowed for the selection and registration of some rust-resistant material (Rubio et al. 2006).

For lentil and faba bean, simple crossings and single seed descendant selection have also made possible the registration of improved lines with significant levels of resistance to rust incited by U. viciae-fabae that are available to farmers (Idrissi et al. 2012; Rubiales and Khazaei 2022; Sakr et al. 2004). These breeding strategies demonstrate the integration of genetic resistance into cultivars, providing a sustainable approach to manage rust diseases in these legume crops, which have been developed to combat other legumes rusts (Deshmukh et al. 2020; Paul et al. 2010). However, the described conventional breeding approaches to obtain legume resistant varieties against rust are time-consuming and not very efficient for complex resistance traits. To increase the efficiency and speed of breeding programs, precision breeding approaches, based on molecular innovations, have been developed and is been applied to develop additional rust-resistant legume cultivars.

Precision breeding strategies

Genomic technologies, including genome sequencing, resequencing, genetic map**, and diverse omics strategies, are crucial in legume precision breeding. Advances such as next-generation sequencing (NGS) have led to techniques like genoty** by sequencing (GBS), diversity array technology sequencing (DArTseq), RNA-sequencing, and whole-genome sequencing (WGS), significantly improving marker technologies. These have enabled the discovery of numerous single nucleotide polymorphisms (SNPs) closely linked to genes or QTLs controlling rust resistance, enabling faster and more accurate breeding. A compilation of different types of molecular markers closely associated with legume rust resistance genes (previously described in section “Resistance Mechanisms Against Rust”) is presented, which could be beneficial for marker-assisted selection (MAS). These advancements in genomic tools and techniques signify a substantial leap in the ability to understand and manipulate genetic factors underlying rust resistance in legumes, potentially transforming breeding programs and enhancing crop resilience.

Precision breeding in peanuts for rust resistance involves the introgression resistance genes from wild species into cultivated varieties, utilizing MAS for efficient selection (Bertioli et al. 2016). For instance, markers like SSRGO340445 and SSRIPAHM103 have been identified near P. arachidis resistance loci (Varshney et al. 2014) and used in marker-assisted backcrossing to introgress resistance into elite peanut genotypes, enhancing rust resistance in cultivated peanut varieties (Ramakrishnan et al. 2020).

Common bean serves as a prime example of a legume crop where, to date, the highest number of molecular markers associated with with U. appendiculatus resistance has been identified. Decades of dedicated efforts have led to a thorough understanding of different race-specific resistance genes, facilitating the development of various marker-assisted selection (MAS) strategies to improve rust management in the field. The resistance gene Ur-14, present in the Ouro Negro cultivar, has been transferred to offspring using its flanking markers RAPDOXY11 and SCARF10 (Ragagnin et al. 2009). Screening segregating populations with these polymorphic molecular markers in combination with other markers associated with resistance to additional diseases, has enabled the development of elite varieties resistant with multiple disease resistance through a detailed MAS scheme (Ragagnin et al. 2009). Subsequently, the resulting resistant cultivar has been used as a parental donor of Ur-14 in crosses and backcrosses with parental donors of resistance genes Ur-5 (markers RAPDOPF10 and SCARSI19) and Ur-11 (markers RAPDOPAC20 and SCARFSAE19) (Souza et al. 2014) to pyramid diverse rust resistance sources into elite varieties using MAS (Pilet-Nayel et al. 2017; Souza et al. 2014). In this context, the identification of high-quality SNP markers and specific genes will continue to expand knowledge and tools for MAS in common beans, as evidenced by recent GWAS studies (Leitão et al. 2023). These advancements demonstrate the significant progress in the application of genomic tools in breeding for rust resistance in legumes, particularly in enhancing the effectiveness of MAS.

However, the utilization of molecular markers to aid breeding for rust resistance in other legume species has not been widely adopted in breeding programs so far. Various factors contribute to this limitation. A significant limiting factor may be the substantial genetic distance existing between most identified markers and the linked resistance genes/QTLs. Most legume markers listed in Table 2 are not closely linked to rust resistance genes/QTLs. Large cM distances often identified in linkage maps make their application in precision breeding challenging. For instance, the distance of most markers linked to rust resistance identified in lentil exceeded 5 cM, while others even surpassed 10 cM, signifying a considerable gap between the marker and the resistance gene (Kant et al. 2004). Furthermore, not all markers are ideal for MAS. Specifically, OPX-15,760 and OPX-171,075, implicated in resistance to U. viciae fabae in lentils, are RAPD markers which can lead into limitations in their reproducibility and detecting allelic variants among heterozygotes (Jiang 2013). In addition, the exact distance between the (SSR) marker and the gene/QTL or their location on linkage groups/chromosomes, is not always known (Dikshit et al. 2016; Fikru et al. 2014). These challenges highlight the complexity of integrating molecular markers into legume breeding programs for rust resistance and underscore the need for continued research to refine these tools for more effective application.

The identification of SNPs markers saturating the plant genome is gaining widespread acceptance due to their diverse applications in plant breeding and genetics (Hickey et al. 2019). These markers, numbering in the thousands, allow to study the genetic diversity within core collections of plants. The high-resolution data provided by SNPs allows for an in-depth examination of the genetic makeup, including analysis of linkage disequilibrium and population structure, which are essential components in whole-genome studies (Rispail et al. 2023). In the context of GWAS, these SNPs markers are crucial for identifying allelic polymorphisms within genes that are involved in resistance mechanisms. This is particularly significant for enhancing marker-assisted breeding, as it allows for greater precision and efficiency in selecting for traits such as rust resistance in legumes (Susmitha et al. 2023). Moreover, genotyped collections derived from SNP markers are helpful for fitting genomic selection (GS) models. This GS strategy enables the prediction of phenotypes in various unknown plant varieties, considering different agroclimatic environments (Annicchiarico et al. 2019). In legumes, both GWAS and GS studies have played a pivotal role in uncovering new genes linked to disease resistance and their utilization in breeding (Zargar et al. 2015). The markers identified through these studies facilitate the breeding of new varieties with enhanced resistance to diseases like rust.

The effectiveness of these precision breeding methodologies is not limited to legumes. Therefore, the advancements in genomic tools and procedures show a substantial improvement in breeders’ ability to incorporate biotechnological methods into conventional breeding strategies. Post-genomic reverse genetics techniques, such as RNA interference (RNAi) and Targeting Induced Local Lesions IN Genomes (TILLING), are being used to confirm genetic functions and expedite the selection process for desirable traits (Padilla-Roji et al. 2023). Although initially more common in cereals for combating diseases like rust, the application of these techniques in legumes, such as peas, has shown promising results. For instance, allowing the characterization of nodulation trait mutants in peas (Tayeh et al. 2015). In M. truncatula, suppressing pathogenesis-related gene expression through gene silencing of a yeast protein MtSTP13 increased susceptibility to powdery mildew, while its transient overexpression enhances resistance against the disease in peas (Gupta et al. 2021). This expansion of genomic and molecular techniques in plant breeding, particularly in disease resistance, marks significant progress adaptable to other legumes. The ability to accurately quantify genetic diversity, identify disease-resistant genes, and predict phenotypic outcomes across different environments accelerates the development of improved crop varieties, ultimately enhancing crop resilience, crop productivity and an efficient disease management.

Similarly, the latest advancements in genetic transformation including CRISPR/Cas9 and single-guided RNA sequence (sgRNA) have achieved ambitious goals in rust resistance, though currently limited to cereals (Hickey et al. 2019). Using the CRISPR/Cas9 technique, it has been possible to silence Ca + transporter genes involved in disease resistance, leading to the development of wheat varieties resistant to Puccinia striiformis f. sp. tritici (He et al. 2023). This breakthrough demonstrates the potential of advanced gene-editing techniques in enhancing rust resistance in crops. Its successful application in wheat suggests promising prospects for its use in legumes, where some advances have recently been made in pea, cowpea, and common bean crops (de Koning et al. 2023; Ji et al. 2019; Li et al. 2023). The ability to precisely edit genes linked to rust resistance can significantly accelerate the breeding of resistant varieties, potentially revolutionizing the management of rust diseases in legume crops.

Conclusions and future prospects

Rust diseases caused by a wide range of causal agents present substantial challenges in legumes. The management of these diseases demands integrated approaches that are both environmentally sustainable and cost-effective. Efforts to enhance rust resistance in legumes have led to the development of cultivars with varying levels of incomplete resistance. Despite these progresses, available resistance sources against rust are still limited, and current screening methods remains laborious and time-consuming. This highlights the necessity for more accurate phenoty**, achievable through the integration of novel, high-throughput phenoty** platforms.

Significant strides in genomic technologies, including genome sequencing and genome-wide discovery of marker-trait associations, are essential in enhancing legume breeding strategies. The availability of optimized reference genomes in several legumes such as pea (Kreplak et al. 2019; Yang et al. 2022), common bean (Schmutz et al. 2014) or faba bean (Jayakodi et al. 2023) are proving invaluable at the molecular level. These resources aid in precise breeding and in unravelling complex trait genetics for better genetic gains. Collaboration among breeding programs is also crucial to share diverse genetic materials, facilitating the gene pyramiding of useful traits and implementing various hybridization regimes to stabilize resistance genes.

The complex nature of rust diseases in legumes requires multidisciplinary approaches that employ both biological knowledge and policy directives to enhance environmental sustainability and food security. Exploring resistance in lesser known but resilient legume species could provide valuable insights for breeding major legume crops. The resistant rust gene transferred from pigeon pea to soybean serve as a successful example (Kawashima et al. 2016). Therefore, investigating wild relative species for rust resistance has also the potential to significantly contribute to the improvement of other major legume crops.