Secondary Metabolites, Other Prospective Substances, and Alternative Approaches That Could Promote Resistance against Phytophthora infestans

Abstract

Potato (Solanum tuberosum) is a valuable staple crop that provides nutrition for a large part of the human population around the world. However, the domestication process reduced its resistance to pests and pathogens. Phytophthora infestans, the causal agent of late blight disease, is the most destructive pathogen of potato plants. Considerable efforts have been made to develop late blight-resistant potato cultivars, but the success has been limited and present-day potato production requires the extensive use of fungicides. In this review, we summarize known sources of late blight resistance and obstacles in P. infestans control. We outline the problematic aspects of chemical treatment, the possible use of biological control, and available resources of natural resistance in wild Solanum accessions. We focus on prospective putative markers of resistance that are often overlooked in genome-centered studies, including secondary metabolites from alkaloid, phenylpropanoid, and terpenoid classes, lipids, proteins, and peptides. We discuss the suitability of these molecules for marker-assisted selection and the possibility of increasing the speed of conventional breeding of more resilient cultivars.

1. Introduction

Archeological evidence indicates that wild potato species were part of the human diet more than 10,000 years ago [1], and potato domestication is believed to date back around 8000 years to the Andean region of the present-day states of Peru and Bolivia [2]. The commonly known potato (Solanum tuberosum) was introduced to Europe in the late 16th century. It was rapidly spread across the world by European colonial powers, becoming a major staple in many regions (Figure 1). Potato cultivation is said to have contributed to the industrialization of Europe by nourishing and boosting the population in the 17th and 18th centuries, and the failure of potato production caused by Phytophthora infestans led to the infamous Great Famine in Ireland [3]. Today, annual potato production has reached 360 million tons (Food and Agriculture Organization of the United Nations; https://www.fao.org/; accessed on 10 December 2022). Potato is a more efficient crop than wheat, with higher yields and lower demands on water consumption [4]. However, a part of its domestication included a decrease in protective toxic compounds such as glycoalkaloids [5]. That has reduced its resistance to pests and pathogens, and potato is one of the most fungicide-dependent crops in the world [6]. It is affected by insects, nematodes, fungi, oomycetes, bacteria, and viruses, resulting in an estimated annual yield loss of about 20% [7]. P. infestans (late blight), Alternaria solani (early blight), Streptomyces scabies (common scab), Rhizoctonia solani (black scurf), Ralstonia solanacearum (bacterial wilt), and PVY^N (tuber necrosis) remain the most important reemerging pathogens in the potato industry worldwide [8].

It took more than a century to establish potatoes as a crop of global importance in Europe. In contrast, the spread of late blight occurred within decades (Figure 1), and only a few recorded crop failures have been as devastating as those caused by this pathogen in the 1840s. The geographic origin of P. infestans is disputed, with some studies indicating a Mexican center of origin [11] and some providing evidence for a region in the Andean highlands [12]. Interestingly, the original FAM-1 genotype went extinct and has been replaced by more aggressive lineages [9]. The rapid evolution of this pathogen remains a major threat to global food security, and significant yield loss results in annual economic losses estimated at billions of euros [13]. This review summarizes current strategies and limitations for controlling potato late blight. It provides insight into P. infestans resistance and discusses novel approaches and molecular markers that could be utilized to speed up the breeding process of new resistant cultivars.

2. Phytophthora infestans

Phytophthora (from the Greek for plant-destroyer) is a globally distributed genus of fungus-like oomycetes that cause agricultural and ecological plant diseases [14]. It includes more than 200 taxa distributed across twelve clades, and pathogens with a significant negative impact on agricultural production are found in several of these. For example, P. infestans belongs to clade 1, together with P. parasitica, P. nicotianae, and P. cactorum. The second clade includes P. capsici, and the representants of clade 4 are the pathogens P. megakarya and P. palmivora. Finally, P. sojae and P. cinnamomi are found in clade 7 [15]. P. infestans was first described by Heinrich Anton de Bary, who studied the potato disease responsible for the Great Irish Famine [16]. It is an obligate hemibiotrophic pathogen that attacks living tissues (leaves, stems, tubers) of the Solanaceae plant family [17]. Most infections during a season are initiated by rapid asexual reproduction, and P. infestans can be spread aerially through asexual sporangia. Spores enter leaves and stems through natural openings (e.g., stomata), buds, wounds, or by direct cuticle and epidermis disruption, using a slicing mechanism called a naifu invasion [18]. The long-distance movement and spreading of the pathogen are mainly anthropomorphic, due to global trade and the unintended transport of infected tubers for use as seed potatoes [19]. P. infestans may overwinter in tubers which then serve as the primary inoculum in the field. In parallel, contact between infected tubers and healthy tubers in storage promotes infection spread, and the careful removal of infected tubers before storage, forced air ventilation, controlled temperature and humidity, and disinfectants are commonly used to protect the harvest and mitigate disease transmission [20]. P. infestans takes up nutrients from living plant tissues until the terminal phase of infection occurs. This phase is associated with host necrosis and pathogen sporulation, which can occur as early as three days after leaf infection [21]. Rapid sporangia production and its long-distance dispersion by the wind result in the ability of P. infestans to destroy unprotected crop vegetation in a few weeks [22,23]. Asexual zoospores are essential for population growth. However, due to the spread of alternative mating types (A1 and A2), the co-occurrence of both mating types and sexual reproduction is also possible, presenting a risk of the emergence of more aggressive P. infestans genotypes [24,25,26]. The mating in P. infestans leads to the production of thick-walled oospores that serve as survival structures and additional sources of inoculum [27]. Sexual reproduction is area-specific and seems to be preferred under stress conditions. Interestingly, the most aggressive known P. infestans isolates are triploid clonal lineages US-1 and 13_A2, and under stress the triploid genotype can change to a diploid one [28].

3. Resistance to P. infestans

3.1. Plant Innate Immunity

Plants have evolved several layers of defense to resist pathogen attacks. In addition to a passive defense based on physical barriers and chemical composition, plants have evolved active defense mechanisms. In the classical model of plant–pathogen interaction (Figure 2), pattern-triggered immunity (PTI) is activated by recognizing molecular patterns associated with pathogens or damage. Phytophthora produces several unique compounds that are recognized by plant receptors, including cysteine-rich proteins elicitins, eicosapolyenoic acids, and ß-glucans released from the cell wall of oomycetes by plant glucanases [29]. Successful PTI leads to ROS burst and hypersensitive response, the expression of defense genes, callose deposition, and the accumulation of protective secondary metabolites. P. infestans produces effectors that block PTI responses or stimulate plant susceptibility factors [21,30,31,32]. In turn, plants have evolved resistance proteins (R-proteins) that facilitate effector recognition and effector-triggered immunity (ETI) activation, often resulting in a hypersensitive reaction [33]. The PTI and the ETI are associated with so-called qualitative (vertical) and quantitative (horizontal) resistance, respectively. Qualitative resistance is, by definition, based on a single major resistance gene (R gene), providing efficient protection against a pathogen genotype producing specific protein. The durability of R gene-based resistance is limited by multiple obstacles, including the sheer number of effector proteins that are evolving and increasing the pathogen’s ability to escape recognition mediated by R proteins and establish a successful infection. The best-described virulent effectors are RXLR (conserved Arg-Xaa-Leu-Arg motifs in their N-terminal sequence) and CRN proteins (CRinkle and Necrosis phenotype). These proteins are secreted by Phytophthora and are destined to be translocated and function inside host cells [34]. P. infestans encodes approximately 560 and 190 RXLR and CRN effectors, respectively, and of these roughly one half are assumed to be important for the infection [35,36,37]. The effectors that are recognized by R proteins and trigger ETI are referred to as avirulence proteins, and a mutation in the corresponding genes and gene silencing enable Phytophthora to escape the recognition [38]. P. infestans has an enormous capacity to adapt to plant defense mechanisms. Continuous and rapid changes in P. infestans populations caused by sexual recombination, persistent oospores, rapid asexual reproduction, genome plasticity, and international migration have significant implications for the rapid evolution of pathogen virulence, unbalancing the arms race in its favor. Consequently, cultivated potatoes suffer from reemerging late blight epidemics [22].

In contrast to qualitative, quantitative resistance is the combined result of interactions between multiple genes referred to as quantitative trait loci (QTLs). These genes/proteins play only a minor role in defense, but the resulting additive effect of a large number of components with minor individual influences promotes resistance. Plants do not show complete resistance, but the process is more robust and is usually not pathogen-specific [39,40,41,42]. The number of putative QTLs identified is steadily increasing [43,44,45,46], but the molecular mechanisms underlying the resulting resistance are far from understood. An example of a known mechanism of quantitative resistance to P. infestans is cell wall thickening due to the deposition of hydroxycinnamic acid amides, flavonoids, and alkaloids [47].

It should be noted that the differences between quantitative and qualitative resistance are not as concrete as initially thought. Many QTLs contain clusters of known R gene homologs that seem to be involved in the QTL effect (e.g., [48]). For example, several R genes identified in wild potato species confer broad-spectrum resistance [49].

Figure 2. The simplified model of potato immune responses to P. infestans. PAMP, pathogen-associated molecular pattern; DAMP, damage-associated molecular pattern; E, pathogen’s effector; R, plant’s R protein; PRR, pattern-recognition receptor; PTI, pattern-triggered immunity; ETI, effector-triggered immunity; RRP, resistance-related proteins; RRM, resistance-related metabolites. References: 1 [31], 2 [50], 3 [29], 4 [51], 5 [30], 6 [52].

Figure 2. The simplified model of potato immune responses to P. infestans. PAMP, pathogen-associated molecular pattern; DAMP, damage-associated molecular pattern; E, pathogen’s effector; R, plant’s R protein; PRR, pattern-recognition receptor; PTI, pattern-triggered immunity; ETI, effector-triggered immunity; RRP, resistance-related proteins; RRM, resistance-related metabolites. References: 1 [31], 2 [50], 3 [29], 4 [51], 5 [30], 6 [52].

3.2. Chemical Treatment

Late blight is predominantly managed by the continuous use of fungicides, and experimental evidence indicates that 20–60% of annual production would be lost without it [53]. The use of chemicals against P. infestans dates to the late 19th century, with the application of the Bordeaux mixture (CuSO₄, CaO) discovered by Pierre-Marie-Alexis Millardet in 1882. Its application was gradually replaced by the next generations of fungicides because copper is persistent in the environment and phytotoxic, but it is still used in organic agriculture [54]. The European register in 2022 listed 37 fungicides and fungicidal mixtures, including both preventative and curative compounds (EuroBlight, https://agro.au.dk/forskning/internationale-platforme/euroblight/, accessed on 10 December 2022), and some commonly used compounds are listed in Figure 3. Traditional treatment depends on preventive fungicides applied regularly during the growing season, and the treatment period depends on weather conditions and cultivar resistance, as well as fungicide characteristics and efficacy. Readers interested in these aspects and the benefits of precision agriculture in the fight against late blight are referred to the report published by Yangxuan Liu et al. (2017) [55]. Preventive fungicides must be present before infection, and only a selected few provide systemic protection [56]. More importantly, P. infestans has already evolved resistance to some fungicides [57], and strict resistance management measures are required. The overexposure of target populations to single-site fungicides should be avoided [58]. A less effective but more environmentally friendly approach than a conventional fungicide is treatment with phosphite (e.g., KPO₃). Interestingly, its application can reduce the dosage of fungicides without compromising protection against late potato blight [59,60,61]. A similar effect was found with the direct application of phosphorous acid, when a significant reduction in the severity of the disease was observed even at half the recommended concentration of fungicides [62].

3.3. Biological Control

The risk of acquired tolerance to fungicides and public demand to limit pesticide applications have driven the search for alternative chemicals of natural origin and natural enemies of the pathogen. The search has provided a large number of candidates, with some notable examples listed in

Table 1. Representative biocontrol agents tested for late blight disease management.

Examples of Biocontrol Agents	Organism	Effect	References
Extracts	Xanthium strumarium, Lauris nobilis, Salvia officinalis, Styrax officinalis	Mycelial growth inhibition	[73]
	Solanum habrochaites	Mycelial growth inhibition, reduced disease progression	[75]
	Willaertia magna C2c Maky	Disease reduction	[78]
	Trichoderma virens	reduced disease progression	[72]
Bacteria	Pseudomonas strains isolated from the rhizosphere and shoots of potato	Reduced disease progression	[79]
	Bacillus subtilis	Reduced disease progression	[80]
	Myxococcus fulvus	Reduced disease progression	[81]
	Bacillus velezensis	Mycelial growth inhibition, improved resistance	[77,82]
Fungi	Fungal endophytes isolated from Solanum spp.	Mycelial growth inhibition	[69]
	Fusarium oxysporum	Induction of systemic resistance	[83]
	Fungal endophytes isolated from Espeletia spp.	Mycelial growth inhibition	[68]
	Endophytes Phoma eupatorii	Mycelial growth inhibition, infection prevention	[70]

. Promising targets are endophytes, microorganisms growing within plants without causing apparent disease symptoms in their host [68,69,70]. The protective effect of these microbes can originate from different mechanisms, including a simple competition for nutrients and space, the production of antifungal compounds, and the priming of the host’s defense mechanisms [71]. The direct application of living microbes could have undesirable side effects on the plant host, the microflora, and the soil environment. This can be avoided by the application of cell-free extracts [72,73]. On the other hand, the metabolome of endophytes varies depending on their environment. Therefore, an in vitro or ex planta production is unlikely to produce an identical composition of compounds to that found in the natural environment [74]. In addition to microbial extracts, a recent study showed that extracts of plants resistant to Phytophthora could be effective protectants [75]. Interestingly, despite years of experimental evidence and numerous candidate biological agents, the transfer of in vitro and greenhouse results to the field has not been very successful for late blight control. This is not surprising because biological control is dependent on the environment, and both abiotic factors and indigenous microbes can suppress its protective effects by limiting the growth of the biological agent or affecting its effect [76]. Part of that problem could possibly be circumvented by exploiting natural potato endophytes. A recent study isolated more than 200 endophytes from the healthy roots of field-grown potatoes and showed that a significant portion of these microbes manifested anti-oomycete activity [77].

3.4. Resistant Cultivars

3.4.1. Natural Resistance

The impact of late blight on yield can be avoided by using varieties that are less sensitive to P. infestans or by very early bulking to escape late blight infection. However, as illustrated in Figure 4, resistant cultivars represent only a minority in the list of registered cultivars, and public demand requires the production of many traditional cultivars that are not resistant. An extensive resource of resistance is hidden in more than 100 wild potato species, with late blight resistance developed during the ages of host and pathogen coevolution [84,85,86,87,88]. The number of recognized species resistant to Phytophthora has been steadily growing, and both Solanum genotypes with foliar resistance and tuber resistance to P. infestans have been identified [89,90,91,92,93]. Wild germplasms have been exploited in breeding processes for a long time, including S. demissum, S. bulbocastanum, S. stoloniferum, and S. verrucosum (as reviewed in [94]), and more than 70 Rpi genes have been identified and mapped in 32 Solanum species [95]. However, only a minority of identified species resistant to P. infestans can be directly used for resistance breeding in tetraploid potato cultivars (2n = 4x = 48) without the manipulation of ploidy by chromosome doubling, unreduced gametes, bridge crosses, somatic fusions, or other means to circumvent hybridization barriers [96].

3.4.2. Traditional Breeding

The introgression of durable resistance to Phytophthora into cultivated potatoes is a time-intensive process (Figure 5) and often ends in disappointment. The limiting factors in late blight resistance breeding are not only the long breeding cycles and the genome of cultivars that are highly heterozygous tetraploid plants. In fact, most of the resistant cultivars obtained had insufficient table or processing quality for commercial success, or resistance was soon lost due to the rapid evolution of the pathogen [98,99]. For example, a very promising introgression of an R gene from S. demissum into cultivated potatoes started in the early twentieth century and took several decades. The resulting cultivar was soon found to be ineffective [100], and decades-long breeding procedures failed in only a few growing seasons when new virulent races of P. infestans evaded resistance. The Sárpo Mira cultivar is one of the few potato cultivars that has been reported to retain resistance in the field for more than a decade. Its genetic origin has not been publicly disclosed, but its genome contains a set of at least five different R genes that confer qualitative and quantitative resistance to late blight [101,102,103]. It has been used in studies of late blight resistance and as a source of resistance in conventional breeding [102,104]. On average, the traditional potato breeding process takes more than ten years from the initial crossing (Figure 5) to obtaining a new cultivar, but the process can take considerably more time due to a laborious selection process, backcrossing, and the elimination of undesirable quality traits [105]. The introgression of a single R gene from the wild species S. bulbocastanum into the Bionica and Toluca potato cultivars took an astonishing 46 years [106]. Investment in conventional breeding and multilateral collaboration can significantly improve the throughput of the process, as exemplified by the Bioimpuls project that has been running since 2009 and has provided a constant flow of breeding clones with introgressed resistance genes [107].

3.4.3. New Breeding Technologies for Potato Improvement

Modern genetic engineering techniques are more efficient and faster in introducing resistance genes into susceptible cultivars than traditional breeding. These techniques have been successfully used, especially in the pyramiding (stacking) of the R genes, to improve both the durability and the level of resistance [39,95,108,109,110,111,112,113,114,115,116,117]. The GMO cultivars were successful in field tests [106,116]. However, despite using only genes from wild Solanum species crossable with cultivated potato varieties (a cisgenic approach), the global ostracization of GMOs and legal restrictions have significantly limited the use of these GMO cultivars. Interestingly, in addition to stacking resistance genes for more durable resistance, transformation allows a successful late blight control based on the gene silencing of susceptibility genes [118,119,120].

3.4.4. Somatic Hybridization

The cultivated potato was one of the first crops successfully cultured in vitro and used to obtain somatic hybrids [121]. Somatic hybridization via protoplast electrofusion enables the fast and robust production of potato breeding material. This method of genetic manipulation is not subject to GMO legislation, and the resulting plants are not considered genetically modified (Directive 2001/18/EC—Annex 1B). Somatic hybrids are generally not suitable for cultivar trials but could be very useful as pre-breeding lines, especially to overcome hybridization barriers, as documented in recent literature [111,122,123,124,125,126].

4. Marker-Assisted Selection

The breeding process is considerably accelerated by a cost-effective molecular marker-assisted selection (MAS) diagnostic for both qualitative and quantitative traits. The approach to the detection of resistant genotypes is based on identified biomarkers that qualitatively occur only in resistant genotypes or are significantly more abundant in resistant plants (compared to susceptible ones) [127]. MAS applied to preselected crosses in the fourth year of the breeding cycle (Figure 5) could shorten the breeding process by at least three years [87].

4.1. Genome-Based Analyses

Genomic techniques are powerful and robust tools for the detection of resistance markers and are the dominant approach in MAS [128]. DNA markers have been widely used in various breeding processes [45,129,130], and several have been developed to be used in MAS for resistance to late blight [112,131]. These markers are based on dominant R genes and identified QTLs, and the list of these is steadily increasing. For example, the mapping of R genes belonging to the nucleotide-binding and leucine-rich repeat (NB-LRR) family aided by resistance gene enrichment sequencing (RenSeq) identified more than 750 putative R genes in the reference genome of S. tuberosum [132]. A detailed list of identified late-blight resistance genes has recently been summarized in an excellent review article [95], and readers are referred to that source for further information.

4.2. Metabolome-Based Analyses

Genetic markers are associated with a trait of interest. Ideally, the marker originates from the allelic variant of the gene that governs the desired phenotypic effect. However, durable quantitative resistance is controlled by an intricate network of genes and other regulatory elements. Some of these elements are involved in transcription regulation and gene silencing, while others determine biochemical traits and metabolite production. These genes provide an additive effect that is difficult to understand, but the corresponding metabolome signatures are directly measurable and are often described as metabolic QTL (mQTL) [133]. General applications of metabolomics in plant breeding have recently been reviewed [134,135]. Here, we will discuss the potato metabolome and putative targets for late blight resistance breeding. It is estimated that thousands of diverse metabolites are synthesized by potato plants, but only a relatively small part of these are routinely identified and quantified. For example, a recently published database of root, tuber, and banana crops contains data for fewer than 100 metabolites of S. tuberosum [136]. Many recent studies had a similar metabolome coverage in the mid-hundreds of identified metabolites [137,138,139,140]. Despite advances in the sensitivity of the metabolomic methods, the analyses are limited by at least three factors: (i) dynamic concentration ranges and spatiotemporal changes in the metabolome, (ii) the diversity in physicochemical properties of metabolites that necessitates the use of complementary analytical workflows, and (iii) a large number of unidentified metabolites. Untargeted metabolomics analyses allow for the measurement of both known and unknown metabolites. Unknown metabolites (features) are characterized by unique retention time, specific mass, and fragmentation spectra. In theory, these features could be used as mQTLs for MAS but are usually excluded because the underlying mechanisms of the observed phenotypes are lacking.

Identified resistance-related metabolites are classified according to their biosynthesis as phytoanticipins (constitutive production) and pathogenesis-induced phytoalexins [141]. Plant phytoalexins and phytoanticipins comprise metabolites produced in various metabolic pathways, including the shikimate pathway (phenylpropanoids), mevalonate pathway (terpenes), the urea cycle (alkaloids, polyamines), carbohydrate metabolism, and fatty acid metabolism [127].

4.2.1. Alkaloids

Plants produce a wide range of nitrogen-containing secondary metabolites with antimicrobial activity, called alkaloids [142]. Plants in the Solanaceae family produce tropane alkaloids, pyrrolizidine alkaloids, and steroidal glycoalkaloids [143]. These compounds protect against insects, herbivores, and pathogens. The two main and best-described potato alkaloids are α-solanine and α-chaconine and they belong to the group of cholesterol-derived steroidal glycoalkaloids. Tubers from wild potato species commonly contain these glycoalkaloids at concentrations that exceed international health regulations for human consumption [144], and melicopicine, solanidine, α-chaconine, and α-solanine were identified as constitutive metabolites related to resistance to late blight [145,146] (

Table 2. Alkaloid inhibitors of P. infestans. NA, not available.

Class	Name	HMDB
Steroidal saponins	α-Solanine	HMDB0034202
	α-Chaconine	HMDB0039353
Acridone alkaloids	Melicopicine	NA
Solanidines	Solanidine	HMDB0003236
Tropane alkaloid	Scopolamine	HMDB0003573

). Interestingly, several previous studies indicated that glycoalkaloid content per se is not the trait correlated with resistance to late blight [144,147,148]. It seems that only the nonglycosylated glycoalkaloid precursor solanidine is a potent inhibitor of P. infestans, indicating the host-specific adaptation to potato glycoalkaloids [149]. The potential human toxicity of glycoalkaloids has led to guidelines that limit the glycoalkaloid content of new cultivars [150], and a recent study claimed the lowest observed adverse effect level of daily glycoalkaloids consumption at 1 mg per kg of body weight [151]. In summary, solanidine is a prospective marker for genotypes with attenuated glycoalkaloid production, e.g., genotypes with a limited conversion of solanidine to solanine and a total alkaloid content within the safety limits. A recent study showed that tropane alkaloid scopolamine inhibits sporangia germination and viability, and that its application could promote the effects of a chemical pesticide [152]. This alkaloid is found in some members of Solanaceae, but given its infamous reputation, it is unlikely that it would find a large-scale application in agriculture.

4.2.2. Phenylpropanoids

Phenylpropanoids represent a large class of secondary plant metabolites derived from aromatic amino acids phenylalanine and tyrosine. It is estimated that variations of the substituents on the benzene ring and the position of the propenyl double bond result in more than 8000 different phenylpropanoid metabolites [153]. One of these compounds is salicylic acid, a plant hormone well known for its role in plant defense mechanisms. The salicylic-acid-deficient potato mutant showed a drastic increase in pathogen growth that was correlated with compromised callose formation and reduced early defense gene expression [154]. However, a higher level of salicylic acid is not correlated with resistance to late blight [155]. Hydroxycinnamic acid and other phenolic acids exhibit antimicrobial activity, and cell wall-bound phenolics contribute to cell wall strengthening and restrict fungal penetration [156]. Many phenylpropanoid compounds are more abundant in resistant genotypes or induced in response to P. infestans, and a recent study showed that the phenylpropanoid pool composition could correlate with resistance [157]. Candidate markers of resistance include coumarins, derivatives of hydroxycinnamic acid, flavonoids, and quinic acid and derivatives (

Table 3. Putative phenylpropanoid markers of late blight resistance. The listed compounds were found in at least two of the referenced studies [47,138,145,146,157,159,160,161,162,163,164]. HMDB, The Human Metabolome Database metabolite annotations.

Class	Name	HMDB
Coumarins	4-Coumaryl alcohol	HMDB0003654
	Scopolin	HMDB0303366
	Scopoletin	HMDB0034344
Catechols	Paucine	HMDB0029876
Flavonoids	Rutin	HMDB0003249
Hydroxycinnamic acids and derivatives	1-O-Feruloyl-β-D-glucose	HMDB0302219
	1-O-Sinapoyl-β-D-glucose	HMDB0302379
	Caffeic acid 3-glucoside	HMDB0303040
	Ferulic acid	HMDB0000954
	N-cis-Feruloyltyramine	HMDB0036381
	Subaphylline	HMDB0033463
Quinic acids and derivatives	5-O-Feruloylquinic acid	HMDB0240478
	p-Coumaroyl quinic acid	HMDB0301709
	Chlorogenic acid	HMDB0003164
	Quinic acid	HMDB0003072

). The validity of these putative markers is yet to be tested, but at least some of these compounds show the inhibition of pathogen growth at physiological concentrations [158].

4.2.3. Terpenoids

Terpenoids are one of the largest groups of secondary plant metabolites, with diverse functions in plant growth and development. Volatile terpenoids are involved in biotic stress responses, and their role in repelling pests and attracting herbivore predators is well known [165]. Antimicrobial activity has been reported for volatile and non-volatile terpenoids, and late blight symptoms occur in parallel with a reduction in the expression of a gene that encodes an enzyme that catalyzes the initial step of isoprenoid biosynthesis [166]. An induced accumulation of different classes of terpenoids has been reported in numerous plant-Phytophthora interactions [146,162,167,168], and the known isoprenoids related to late blight resistance are phytuberin, rishitin, and its precursors lubimin and solavetivone [169,170] (

Table 4. Terpenoids that inhibit P. infestans. HMDB, The Human Metabolome Database metabolite annotations; NA, annotations not available.

Class	Name	HMDB
Sesquiterpenoids	Phytuberin	HMDB0035754
	Rishitin	HMDB0035593
	Lubimin	NA
	Solavetivone	HMDB0035657
Saponins	Neoindioside D	NA
	Protoneodioscin
	Barogenin-solatrioside
	Barogenin-chacotrioside

). These compounds inhibit Phytophthora growth but were found to be ineffective when applied to leaf discs or used as protectant sprays [169]. The reason could be toxicity, which is not limited to the pathogen. Plants actively detoxify these phytoalexins through cytochrome P450 [171], and their natural occurrence occurs predominantly in tubers. Putative terpenoid resistance markers suitable for MAS are recently discovered steroidal saponins. Four compounds were identified in the leaves and tubers of potato cultivars inoculated with P. infestans: neoindioside D, protoneodioscin, barogenin-solatrioside, and barogenin-chacotrioside. All of these saponins showed a high anti-oomycete activity with IC₅₀ in the micromolar range [172].

4.2.4. Polyamines

Polyamines are aliphatic compounds that contain two or more amino groups with a positive charge at physiological pH. Polyamine biosynthesis starts with arginine and its conversion to polyamine precursor ornithine or agmatine [173]. The accumulation of putrescine was reported in response to P. palmivora [174] and P. infestans [175]. However, polyamines are ubiquitous molecules involved in various processes, such as growth and development regulation, response to phytohormones, and abiotic stress [176,177,178]. Thus, it is unlikely that polyamines could be suitable MAS candidates.

4.2.5. Lipidome

Lipidome has a critical role in plant biotic interaction [179]. Free fatty acid levels increase with pathogen attack, affecting the composition and fluidity of lipids in the plant membrane or the production of signals derived from fatty acids, including oxylipins [180,181,182]. Early studies showed that oxylipins colneleic and colnelenic acid accumulate more rapidly in a potato variety resistant to late blight and inhibit the growth of mycelial P. infestans [183]. However, a later study did not find any correlation between P. infestans resistance levels and oxylipin synthesis rates or concentration [184]. However, oleic and linoleic acid (oxylipin precursors) were reported to be involved in the defense against fungal, oomycete, and bacterial infections [167,185,186], and were suggested as putative late blight resistance markers [146]. It has been proposed that oil bodies may mediate defense against microbes, especially in senescent leaves [187]. This is in line with the significant accumulation of triglycerides reported in the resistant wild potato genotype S. pinnatisectum [138]. Lipids are also utilized as antibiofilm. The leaf surface of the resistant genotype S. bulbocastanum is covered with heptadecenoyl-lysophosphatidylcholine (LPC 17:1) that inhibits the germination of the P. infestans spore and mycelial growth in vitro [163]. This metabolite was not found on the leaf of S. tuberosum and could be a suitable candidate for MAS.

4.2.6. Volatiles and Other Compounds of Interest

Plants continuously release volatile compounds that facilitate communication and interaction with the environment. The amount of emitted volatiles significantly increases in response to damage, and it is higher under attack by hemibiotrophic or necrotrophic pathogens [188]. The emitted volatile compounds include derivatives of C6-aldehydes, such as Z-3-hexenal, Z-3-hexenyl acetate, and (Z)-3-hexen-1-ol. Recently, pretreatment with Z-3-hexenyl acetate was shown to delay the onset of P. infestans infection and inhibit the intensity of the sporulation [189]. Therefore, these volatile compounds are both markers of biotic stress and candidate MASs for the breeding of more resilient genotypes. An additional candidate for MAS is cysteamine, which accumulates in the resistant wild potato genotype [138]. This simple alkylthiol exhibits antimicrobial activity, and its precursor cystamine inhibits P. infestans growth [190,191] (

Table 5. Other compounds of interest that inhibit P. infestans. HMDB, The Human Metabolome Database metabolite annotations; LM ID, LIPID Metabolites And Pathways Strategy identifier.

Class	Name	HMDB/LM ID
Glycerophospholipid	Heptadecenoyl-lysophosphatidylcholine (LPC 17:1)	LMGP01050002
Acetate ester	Z-3-Hexenyl acetate	HMDB0040215
Dialkyldisulfide	Cystamine	HMDB0250701
Alkylthiol	Cysteamine	HMDB0002991

).

4.3. Proteome and Peptide-Based Analyses

The successful introduction of protein/peptide markers into breeding processes has been hindered by poor accessibility, low throughput, and the higher cost of standard proteome analyses. However, novel techniques have significantly increased the detection limits and processing power of state-of-the-art liquid chromatography-mass spectrometry analyses, reaching deep proteome coverage and a rate of 100 samples per day [192]. In contrast to transcriptomic studies, proteome analysis provides the correct image of the molecular mechanisms underlying the phenotype, including post-translational modifications and real protein abundances that are not easily predictable from transcriptomic data [193]. For example, Ali et al. (2014) reported that only 50% of differentially abundant proteins showed a correlation with gene expression data in S. tuberosum under the P. infestans attack [194]. Thus, quantitative proteomics can generate data that could be directly used for MAS.

4.3.1. Candidate Proteins for MAS

To defeat harmful pathogens, plants employ a complex cocktail of antimicrobial proteins that could be exploited as markers for the early selection of resistant genotypes. The first work that demonstrated that selective monitoring based on tryptic peptides could be a promising technology for marker-assisted selection was published in 2016 [195]. Putative markers identified for P. infestans resistance included two Kunitz-type protease inhibitors, glucan exohydrolase, peroxidase, cystatin-type protease inhibitor, serine carboxypeptidase III, and nonspecific lipid transfer protein (a member of cysteine-rich antimicrobial peptides). Different proteomics studies targeting leaves, tubers, and secreted proteins in the apoplast have extended the list of putative resistance markers. These studies have identified the expected targets (R proteins, osmotins, peroxidases, protease inhibitors, and lipid transfer proteins), as well as transcription factors and multiple defense-related proteins, including glutathione S-transferases, endochitinase, glycosyltransferase, glucosidases, and heat shock proteins 70 [138,194,196,197,198,199,200,201,202,203,204]. However, most of these candidates for MAS need to be validated in dedicated mechanistic studies. A recent study showed that an accumulation of Small G protein StRab5b reduced the lesions on infected potato leaves [205], but it remains to be seen how this alteration impacts the yield and other economically important traits. The most interesting part of the proteome resides in post-translational modifications. However, it is also the least accessible, and there have been only limited attempts to identify putative links between protein modifications and resistance to P. infestans. Two of these exceptions have been protein SUMOylation and protein methylation [201,206], and a recent report showed that the basis of resistance denoted by the avirulence gene Avr8 is in the manipulation of SUMOylation via a deSUMOoylating isopeptidase DeSI2 [207].

4.3.2. Antimicrobial Peptides—Prospective Targets for Enhanced Resistance

Plant antimicrobial peptides are typical for their basic nature, cysteine-rich sequence, and amphipathic design. Many are encoded by a single protein-coding gene, and the resulting precursor proteins are later cleaved and post-translationally modified [208]. Antimicrobial peptides are estimated to account for up to 3% of the Arabidopsis gene repertoire [209]. Plant antimicrobial peptides are usually classified according to their sequence and structural similarity. Antifungal activity has been reported for thionins, defensins, snakins, hevein-like peptides, knottin-type peptides, and α-hairpinins [210,211,212]. Typical potato peptides are snakins that have been shown to mediate protection against a wide range of fungi, bacteria, and yeasts [213,214]. The effectiveness of these endogenous S. tuberosum peptides against P. infestans is unclear, and further research is needed. However, foreign or synthetic antimicrobial peptides have been found to provide resistance against P. infestans, including syringomycin E and syringopeptin 25A from Pseudomonas syringae [205], Stellaria media hevein [215], and the synthetic peptide NoPv1 [216].

5. Conclusions and Future Perspectives

P. infestans is a reemerging potato pathogen that causes significant economic losses worldwide. A considerable amount of work and time has been spent to develop viable late blight management, but the prevailing potato cultivars are susceptible to the pathogen and fully dependent on regular fungicide application. However, the fast-evolving P. infestans has already gained resistance to some fungicides and seems to always be one step ahead of us. Furthermore, the extensive application of fungicides destroys our environment and is not sustainable. Some studies indicate that RNA interference and gene silencing could be the future of protection against P. infestans (reviewed in [217]). It is a publicly more acceptable approach than GMO, but the cost-effectiveness and stability under field conditions remain a challenge [218]. To fight this pathogen, we need new tools that could be found in biological control agents or antimicrobial peptides, or by searching the genome, proteome, and metabolome of wild Solanum species. The available data indicate that the dosage of the fungicide could be decreased by combining it with more environmentally friendly substances. However, the identification of factors such as dosage, application methods, compatibility with the given fungicide, and persistence in the field is needed. We also need to improve our plant breeding techniques, which are very slow and need assistance from molecular approaches. Marker-assisted selection can speed up the process, but it is still limited by the relatively high costs of the analyses. The most time-efficient screening would require analyses of samples from the first field experiment (Figure 5). However, despite the available automatization and sample pooling, 10,000 samples still represent a daunting and expensive task for ‘omics’ analyses. The problem is also due to our limited knowledge of resistance mechanisms and their interaction with abiotic factors. These interactions could be critical, as illustrated in the regulation of RWP-RK transcription factors in soybean infected with P. sojae [219]. As illustrated in this review, many compounds that accumulate in response to infection or are more abundant in resistant genotypes are not directly responsible for the observed phenotype, and more research is needed to identify optimal targets for MAS. Lastly, lifting the GMO restrictions imposed by some governments would significantly broaden our horizons.