Introduction

Biological coevolution is “the change of a biological object triggered by the change of a related object” (Yip et al. 2008) implying that coevolution can occur at multiple levels of biology (i.e., microscopic—changes between amino acids in a protein or nucleotides in genes—or macroscopic—covarying traits between different species in an environment).

In coevolutionary relationships, each member of the association exerts selective pressures on the other, thereby affecting each others’ evolution. Species-level coevolution, during host–pest interactions, implies that the host and pest exert pressures on each other resulting in the combined evolution of the host species and its pests (i.e., host–pest coevolution) (Laine 2006; Whitney and Glover 2013). In most instances, this evolution will be based on a one-on-one interaction (i.e., interactions between predator and prey or host and its parasite/pest) (Laine 2006). Coevolutionary relationships can also be beneficial for both species which then results in a mutualistic symbiosis (i.e., host and its symbiont).

Charles Darwin (1859) briefly described the idea of coevolution in his “On the Origin of Species,” but the concept was only much later thoroughly investigated when taxonomists recognized the immense diversity of secondary compounds (or secondary metabolites) that distinguishes between species and higher taxa of plants compared to the relatively narrow host range of most phytophagous insects. It soon became evident that most secondary compounds were not required for the “primary” functions of plants (i.e., resource acquisition and allocation), and therefore, the postulation that they may be involved in host defense was first put forward. As early as 1920, Brues (1920) suggested, and Dethier (1954) verified, that plant’s secondary chemistry had shaped specialized host associations. In 1959, Fraenkel was the first to compile evidence that most secondary compounds had evolved to defend plants against insects and other natural enemies. However, the conceptualization of the theory was popularly consummated only after Ehrlich and Raven (1964) published their theory on host–herbivore interactions in: “Butterflies and plants: A study in coevolution.”

In their publication, Ehrlich and Raven (1964) suggested that plants evolved novel, highly effective chemical defenses in response to herbivory that enabled them to escape serious damage from most, or all, of their respective associated herbivores. This in turn directed the radiation of plants into diverse species and taxa that, still today, share novel defense systems and similar chemistry. Although they were not able to put forward the specific mechanisms that drive these developments, they suggested that over time, new insect species would colonize close relatives of their hosts and, in so doing, would adapt to plants that share a similar chemistry by shifting from chemically similar, although distantly related, host plants. This suggested that these insects were able to undergo radiation through the exploitation of “empty niches” afforded by a diverse clade of chemically distinctive plants. By doing so, insects underwent adaptive radiation, which lead to the development of new species as described by the authors for butterflies. Ehrlich and Raven (1964) proposed that repetition of such stepwise adaptive radiations through time, in both plant–herbivore and other ecological associations, accounts for a great deal of the observed biological diversity witnessed today (Fig. S1).

An important outcome of the theory by Ehrlich and Raven (1964), and other related studies, was the recognition that plants may adapt to herbivory not only by “resistance” to the pest, but also by “tolerance” of tissue damage (Painter 1951), based on an ability to re-grow or reproduce, by using stored resources (van der Meijden et al. 1988; Núnĕz-Farfán et al. 2007).

Evolution of the host and pest

The host: Triticum aestivum

Grasses originated 55–75 million years ago (MYA) (Kellogg 2001) and presently occupy about 20 % of the physical land area as the dominant taxonomical group (Gill et al. 2004). About 40 MYA, the three major cereals (i.e., rice, maize, and wheat) that serve as staple food for humans diverged from a common ancestor (Fig. 1). Of the three major cereals, common bread or hexaploid wheat (T. aestivum L., 2n = 6x = 42, AABBDD) has the largest genome at 17-gigabase-pair (Gb), containing between 94,000 and 96,000 genes (Brenchley et al. 2012), eightfold larger than maize and 40-fold larger than that of rice (Arumuganathan and Earle 1991). Despite the fact that there exist extensive conservation between the grasses in both gene content and order at low-resolution genetic map level (Gill et al. 2004 for review), comparative genomics have shown that there is little colinearity for disease resistance genes due to the rapid evolution among them (Leister et al. 1998).

Fig. 1
figure 1

The coevolutionary history of humans, cereals, insects, and the insects’ endosymbiont, Buchera aphidicola. The evolutionary history of the Aphidinae is also indicated

Full size image

Since bread wheat consists of three genomes, it is generally accepted that modern bread wheat is the result of a polyploidization event of progenitor species. Firstly, the hybridization of Triticum urartu (AA, 2x) and most likely Aegilops speltoides (BB, 2x) resulted in the formation of Triticum turgidum (AABB, 4x). Thereafter, most likely with human intervention, a hybridization event between T. turgidum (AABB, 4x) and Aegilops tauchii (DD, 2x) formed modern hexaploid wheat (T. aestivum, AABBDD, 6x) (Fig. S2). These classes of cultivated wheat, that is, the one-seeded monococcum (2x), the two-seeded emmer (4x), and the dinkel (6x), were described by botanists as early as 1915. A one-seeded wild relative of monococcum was reported in Greece and Anatolia between 1834 and 1884. The two-seeded wild relative of emmer was discovered by Aaronsohn in 1910 in Lebanon, Syria, Jordan, and Israel (Aaronsohn 1910). It is now generally accepted that wheat cultivation originated in the Euphrates basin as suggested by Candolle in 1886 (Gill et al. 2004).

Interestingly, humans and wheat share a remarkably parallel evolutionary history. Paleontological evidence suggests that when humans and apes diverged from a common ancestor about 4–8 MYA (Wolpoff et al. 1988; Wall et al. 2009), the diploid A, B, and D progenitor species of wheat also diverged from a common ancestor (Fig. 1). Similarly, at nearly the same time (i.e., about 200,000 years ago) when modern humans originated in Africa (Jorde et al. 1998), two diploid grass species hybridized to form polyploid wheat in the Middle East. Humans domesticated wheat 15,000 years ago in the Fertile Crescent (modern-day Iraq and parts of Turkey, Syria, and Iran), marking the dawn of modern civilization (Gill et al. 2004).

Aphids and their endosymbionts

Aphids (Aphididae) are the largest group (~4,700 extant species, Davis 2012) of hemimetabolous, hemipteran, and phloem-feeding insects (Miller et al. 1994). The group consists of three major families, namely the Aphididae (“true” aphids), Adelgidae (adelgids), and the Phylloxeridae (phylloxerans), with most extant species belonging to the family Aphididae (~4,500). This family is believed to have undergone a dramatic radiation between ~5 and 26 MYA (Remaudière and Remaudière 1997), with the development of parthenogenesis and viviparity in the Aphidididae (Davis 2012) (Fig. 1). The subfamily Aphidinae comprised of many members that are important and geographically widespread pests, and highly adapted to environmental conditions and many crop plants (Dixon 1998). Most of the ~4,700 aphid species feed on a narrow range of host plant species, and as such, aphids are significant pests of food globally that pose a major threat to world food security.

Aphids, like other insects, are particularly prone to endosymbiotic associations, and such associations are widespread among members of the orders Diptera (true flies, mosquitoes, gnats, and midges), Homoptera (aphids, whiteflies, mealybugs, psyllids, and cicadas), Blattaria (cockroaches), and Coleoptera (beetles) (Kikuchi 2009; Duron and Hurst 2013). Even though the origin of arthropods, based on first fossil records, is dated at 570 MYA, the estimated divergence of aphids dates back between 150 and 250 MYA (Baumann et al. 1997), suggesting that the original infection of aphids by Buchnera must date back to at least this far (Baumann et al. 1997) (Fig. 1). Following this initial infection, endosymbionts and aphid hosts appear to have diversified in parallel, resulting in the present strains of Buchnera that are associated with the present species of aphid. The resemblance between the phylogenies of host and endosymbiont indicates that the vertical transmission of Buchnera has been maintained since the time of the original infection, with no known transfer of Buchnera between different aphid lineages (Fig. S3). During vertical transmission as in the case of Buchnera, the endosymbiont is passed directly from mother to her offspring (nymphs or eggs), unlike horizontal transmission, where transmission takes place within the same generation directly via air-borne, food-borne, and venereal infection or indirectly via intermediate hosts (Chen et al. 2006). This symbiotic relationship between aphids and Buchnera is an example of obligate mutualism, since it allows aphids the exploitation of nutritionally imbalanced food sources, while Buchnera “lives” in a bilobed structure called a bacteriome, which consists of 60–90 polyploid cells, that is, bacteriocytes. In this relationship, microbial recycling and synthesis of dietary limiting essential amino acids enables aphid population explosions to occur (Baumann et al. 1995, 1997; Moran 1996; Moran et al. 2003, 2005; Sandstrom and Moran 1999; Sandstrom et al. 2000; Telang et al. 1999; Thao et al. 1998).

Diuraphis noxia Kurdjumov (Aphididae), generally known as the Russian wheat aphid, is a small, yellow–green or gray–green elongated (1.4–2.3 mm) phloem-feeding insect with a host preference that includes cereal grasses, favoring mostly wheat and barley. It is suggested that D. noxia coevolved with triticale in the Fertile Crescent, and then distributed from western Asia to Africa and the USA (Fig. 2). The first D. noxia sighting in South Africa was reported in 1978 (Walters et al. 1980), followed by reports from the Texas Panhandle in 1986 (Burd and Burton 1992), where after it spread to Canada (Jones et al. 1989), Chile, and Argentina (Clua et al. 2004). By the mid-1990s, it has successfully spread to all wheat-producing countries except Australia (Basky 2003) with most severe effects on wheat production in South Africa and the USA (Smith et al. 1992; Tolmay et al. 2007; Smith 2009) (Fig. 2 inserts).

Fig. 2
figure 2

Global migration of Diuraphis noxia from its endemic regions in the Fertile Crescent to become an invasive pest species in the Africa and the USA. a Development and migration of D. noxia biotypes in the USA from 1980 (US1) to the development of eight biotypes in 2008 (US2-US8). b Development and migration of South African D. noxia biotypes (SA1, SA2, and SA3) from 1978 to 2008

Full size image

The economic losses attributed to this pest were estimated at around $893 million in the USA from 1987 to 1993 (Morrison and Pears 1998), after which these estimates increased significantly following 2003, with the development of many new D. noxia biotypes in the USA (Botha et al. 2010) (Fig. 2 insert A). New D. noxia biotypes were also reported for South Africa (Tolmay et al. 2007), and presently, there exist three D. noxia biotypes (i.e., SA1, SA2, and SA3) (Jankielsohn 2011) and a highly virulent laboratory-developed (mutant) D. noxia biotype (SAM) (Van Zyl et al. 2005). In South Africa, the wild-type D. noxia is distributed throughout the major wheat-producing areas (i.e., Eastern and South-western Free State), and in 2008, the aphid also migrated to the Western Cape (previously D. noxia free) (Burger et al. unpublished results). It is suggested that drier climatic conditions are one of the factors resulting in the proliferation of D. noxia as these conditions favors the dispersal of this invasive species (Fig. 2 insert A, B).

Although D. noxia has the ability to reproduce sexually and asexually, most instances of new introductions are believed to have happened only by asexual reproduction. In fact, in South Africa, unlike other countries, D. noxia males have never been found. The absence of D. noxia males in many countries eliminates recombination based on sexual reproduction as a source of genetic variation, which in turn explains to some extent the limited observed sequence variation between geographically isolated D. noxia populations (Puterka et al. 1993; Lapitan et al. 2007a; Shufran et al. 2007). However, D. noxia populations from different geographically regions still express varying levels of virulence (Puterka et al. 1992; Smith et al. 1992, Burd et al. 2006). Virulence is the measurement of the ability of a pest to infest a new host. Similarly, populations occurring within the same geographical region may also express different levels of virulence and are then classified as “biotypes.” Biotypes are intraspecific classifications based on biological, rather than morphological characteristics (Shufran et al. 2007), and the biotypic status of D. noxia is thus solely based on the phenotypic response of the plant as a result of the aphid’s feeding (Smith et al. 1992). Therefore, a new D. noxia biotype is a population—independent of geographic location—that is able to injure a cultivated plant containing a specific gene(s), previously resistant to known aphid populations (Basky 2003; Smith 2009). Clearly, the development of new D. noxia biotypes capable of feeding on non-preference host varieties points toward an evolutionary adaptation to biotic stresses preset by humans. This raises the question on how new D. noxia biotypes develop in such short evolutionary time, if cyclic parthenogenesis is the only method of reproduction.

It is suggested that the observed genetic variation in the absence of sexual recombination may be the consequence of recurrent fragmentations of specific chromosomes and chromosome mosaics, such as the female chromosome X as observed in Myzus persicae (Hemiptera: Aphididae) (Monti et al. 2012), which is likely considering the considerable size of the female X chromosome of D. noxia relative to the other chromosomes (Novotná et al. 2011). The genome of D. noxia is packaged into 5 chromosomes with an estimated female genome size of 2C between 0.757 and 0.86 pg (~between 740 and 841 MbFootnote 1) (Doležel and Bartoš 2005) and the 2C for males between 0.661 and 0.70 pg [~between 646 and 684.6 Mb (see foot note 1)] (Doležel and Bartoš 2005). Interestingly, the female X chromosome occupies approximately 35 % of the female haploid genome (1C = 0.43 pg) (Novotná et al. 2011). Preliminary next-generation genome sequencing data from 11 geographically distinct D. noxia populations revealed that the assembled regions of the D. noxia genome are AT rich with an CG content in the order of 25–30 % (Botha et al. 2012), which is comparable to GC contents observed in other sequenced aphid genomes, such as Acryrthosiphon pisum (29.6 %) (The International Aphid Genomics Consortium 2010) and Ahis mellifera (34.8 %) (The Honeybee Genome Sequencing Consortium 2006). These observations may suggest that AT richness within these genomes may be evidence of relaxed GC biases and previous mutational hotspots (Rocha and Feil 2010).

Diuraphis noxia, unlike many other aphids, contains only one endosymbiont, Buchnera aphidicola (Swanevelder et al. 2010). In a recent study, after sequencing the leucine plasmid (>10 kb) of the B. aphidicola genome (~600 kb, Botha et al. 2012), from ten different D. noxia biotypes (i.e., two from South Africa and eight from the USA), a single trinucleotide (CCC) insertion on the leucine plasmid was shown to be the only difference between Buchnera plasmid sequences from these D. noxia biotypes. This nucleotide insertion was shown to be located upstream of the leuA gene (pleuABCD) in an inverted repeat region of biotypes SA1, SAM, USA3, and USA7, and resulted in an increase in length of the biotypes’ leucine plasmid (pleuABCD) (Swanevelder et al. 2010). Although Buchnera from different D. noxia biotypes are known to have altered essential amino acids plasmid copy numbers (Thao et al. 1998; Plague et al. 2003; Latorre et al. 2005), Swanevelder et al. (2010) could not explain the increased expression of the leuA and leuB genes on plasmid copy numbers alone, since low leuA and leuB copy numbers coincided with high gene transcription levels. Modeling data on the stoichiometry of this region suggested that this insertion may increase transcript stability and therefore gene expression (Swanevelder et al. 2010). In a recent study on B. aphidicola strains from Acyrthosiphon pisum and Schizaphis graminum, it was shown that most B. aphidicola genes (e.g., proteins for cellular processing, signaling) present strong evidence for translational robustness (Toft and Fares 2012)—that is, selection favoring protein sequences with increasing robustness to misfolding translational errors (Drummond et al. 2005). However, the same study also showed that genes involved in metabolism have strong selection on nucleotide level, which may be coincidental with a slight selection for translational efficiency to favor expression of genes providing amino acids to the host (Toft and Fares 2012). Whether this holds true for the Buchnera endosymbiont of D. noxia will have to be confirmed. Albeit, the observed nucleotide modification may be a mechanism employed by B. aphidicola of D. noxia to compensate for lower leucine levels, when feeding on previously resistant wheat cultivars, thereby bridging the gap between resistance (incompatible interactions) and susceptibility (compatible interactions), and thus promoting the ensuing adaption of the D. noxia to a new cultivar. Since D. noxia biotype characterization is based on its ability to overcome different host resistances, rather than aphid anatomy and morphology, an endosymbiont mutation that allows feeding on previously resistant cultivars could result in a classifiable “new D. noxia biotype,” and thus a clear example of a coevolutionary adaptation.

Aphids may also willingly or unwillingly recruit other organisms such as phytopathogenic bacteria to enable them to colonize plant hosts by causing disease. In the literature, several of these insect–phytopathogen associations are known, for example Xylella fastidiosa and the sharpshooter leafhopper (Hemiptera, Cicadellidae) (Wayadande et al. 2005; Chatterjee et al. 2008; Fogaça et al. 2010), Pantoea stewardtii and the flea beetle (Esker and Nutter 2002) or pea aphid (Stavrinides et al. 2010), Serratia marcescens and the squash bug (Anasa tritis) (Bruton et al. 1998, 2001; Labbate et al. 2007; Shanks et al. 2007), Erwinia amylovora and the pollinators of apple and pear (Eden-Green and Billing 1974; Spinelli et al. 2005), E. tracheiphila and the cucumber beetles (Ferreira and Boley 1992; Yao et al. 1996), and E. aphidicola and the pea aphid (Gonzalez et al. 2005; Santos et al. 2009). Although some of the intricacies in these interactions are still poorly understood, and the precise coevolutionary process leading to the interaction between the insect and its respective phytobacterium is still unclear, the evidence for coevolution is still undeniable (Nadarasah and Stavrinides 2011). Despite the broad host range of the listed examples on insect–phytopathogenic bacteria associations, the disease symptoms are in many instances remarkably similar to those observed in susceptible wheat after D. noxia feeding, that is induced oxidative stress and formation of necrotic lesions, discoloration of the leaf (chlorosis), leaf rolling and/or wilting, and, eventually under extreme feeding, death of the host (Fouché et al. 1984; Botha et al. 2005, 2006). Recent evidence from studies on the D. noxia secretome (Van Eck 2011) and genome (Botha et al. 2012) revealed the presence of peptides and genomic sequence from bacterial origin (e.g., Erwinia aphidicola), raising the possibility that phytobacteria may also be a contributing factor in the observed complexities in the D. noxia–wheat coevolution.

Evolution of host defense

Detection of invasion by an attacker happens through various effectors such as pathogen-associated molecular patterns (PAMPs) (Chrisholm et al. 2006), microbe-associated molecular patterns (MAMPs) (Boller and Felix 2009), viral coat proteins or herbivore-associated molecular patterns (HAMPs) (Mithöfer and Boland 2008), or insect salivary effectors (Lapitan et al. 2007b; Bos et al. 2010). Once invasion by an attacker has been recognized, the host must balance competing demands for metabolic resources between defense, cellular maintenance, growth, and reproduction (Berger et al. 2007a, b). Plant defense can be costly in terms of plant growth and fitness (Tian et al. 2003; Zavala and Baldwin 2004). In addition to the mobilization of an array of defensive strategies (i.e., upregulation of a suite of defense response genes and production of chemical defense responses), the plant must cope with reduced effective biomass, as well as a decline in photosynthetic capacity in the remaining leaf tissue (Haile et al. 1999; Zangerl et al. 2002; Bilgin et al. 2008, 2010; Nabity et al. 2009). The examination of gas-exchange responses in three wheat lines in response to D. noxia feeding suggested significant physiological costs especially in varieties with antibiotic defense, such as a reduction in chlorophyll fluorescence and photosynthetic rates. Unlike the tolerant line PI 262660 that showed recovery from the injury within a week after removal of the aphid, the antibiotic variety PI 137739 did not express the gradual photosynthetic compensation as observed in the tolerant line (Haile et al. 1999). Although there are few examples of compensatory stimulation of photosynthesis (Botha et al. 2011), most reports suggest that a decline in photosynthetic capacity is inevitable and this may represent the “hidden fitness costs” to defense (Fouché et al. 1984; Haile et al. 1999; Zangerl and Berenbaum 2003, Zangerl et al. 2002, 2003; Heng-Moss et al. 2003; Botha et al. 2005; Bilgin et al. 2008, 2010; Nabity et al. 2009).

Phloem-feeding insects, like D. noxia, also use a variety of chemical and physical stimuli to recognize a suitable host. Initially, D. noxia recognizes its host with sensory neurons on the antennae. Once the aphid lands on a wheat plant, it searches for an appropriate feeding site utilizing a surface scan using receptors on the proboscis to detect a vein (where aphid feeding is preferable) (Dixon 1998). Host recognition is confirmed with a drop of saliva onto the cuticle surface which dissolves the cuticle and this dissolved material is sensed by a chemoreceptor on the labium tip. When the plant is recognized as a suitable host, penetration commences (Srivastava 1987) with probing occurring mostly intracellularly until the phloem sieves are found, after which stylet penetration can occur (Pollard 1973; Tjallingii 2006; Dinant et al. 2010). With penetration, D. noxia secretes two types of saliva at the feeding site along the stylet path. The first is a rapid gelling, sheath saliva, that consist of proteins, phospholipids, and conjugated carbohydrates. These compounds form a protective barrier along the stylet path in order that the stylet does not come into contact with the plant’s apoplast cells. The second type of saliva is a watery, digestive saliva that contains a wide variety of digestive enzymes (i.e., pectinase, cellulases, amylases, proteases, lipases, alkaline and acidic phosphatases, trypsin, and peroxidases) (Miles 1999; Mutti et al. 2008; Cooper et al. 2011; Cui et al. 2012; Nicholson et al. 2012). Secretion of watery saliva seems to be a universal means to prevent clogging inside the capillary food canal (Tjallingii 2006; Will et al. 2007, 2009).

Aphids take advantage of their feeding strategies to avoid and/or deter detection by their host (Walling 2008). They “disguise” themselves through the deliverance of salivary chemicals and/or proteins into the host to influence wound healing (Tjallingii 2006; Will et al. 2007, 2009). Unlike herbivorous insects that inflict major tissue damage during feeding and prime the host for enhanced defense responses (Turlings and Ton 2006; Frost et al. 2008), phloem feeders cause minimal damage and the quantities of herbivore-induced plant volatiles (HIPVs) emitted are low or even undetectable (Du et al. 1998; Turlings et al. 1998; Rodriguez-Saona et al. 2007). Avoiding or lower emission of HPIVs could be beneficial to the aphid since it will result in fewer direct and indirect host defenses (Walling 2008).

These salivary compounds may act as elicitors or effectors in triggering the plants’ primary immune response or effector-triggered immunity (ETI) (Chrisholm et al. 2006). The recognition of the effector is key to ETI activation, and in contrast to vertebrates that have an adaptive immunity system that is based on immunological memory, plants can only rely on their innate immune system in which each individual cell can autonomously mount a defense response (Jones and Dangl 2006). In such system, two layers can be distinguished: one is based on extracellular transmembrane receptors that recognized conserved effectors and induce a relatively weak immune response that halts colonization by most invaders, and the second layer is effective against specialized invaders that breaks through the first layer and is based on highly polymorphic resistance (R) proteins (Takken and Tameling 2009).

Phloem-feeding insects, selectively downregulate the expression of certain photosynthesis-related genes (Heidel and Baldwin 2004; Voelckel and Baldwin 2004, Voelckel et al. 2004; Zhu-Salzman et al. 2004; Qubbaj et al. 2005; Yuan et al. 2005; Botha et al. 2011) and by manipulating the host carbohydrate metabolism, induce a change in carbon flux to their own advantage (Zhu-Salzman et al. 2004; Will and van Bel 2006; Will et al. 2007, 2009; Giordanengo et al. 2010). Herbivory clearly imposes natural selection on plants, sometimes even enforcing habitat specialization. Plants have thus developed varying suites of features to form “defense syndromes” (Argrawal and Fishbein 2006) that could indicate adaptation to particular suites of herbivores and may potentially be dictated by the abiotic environment. A “gene-for-gene” interaction model (Flor 1971, Keen 1990) has been suggested to prevail in resistant wheat plants in response to D. noxia feeding (Botha et al. 2005; Boyko et al. 2006). In this model, host resistance requires the presence of an allelic variant of a plant resistance gene (R-gene) product that recognizes and interacts with the cognate Avr, or effector protein, from the respective pathogen/pest (Keen 1990). Also, since plants are sessile organisms, they have to utilize innovative strategies for defense. One such strategy lies with the “design” of their resistance (R) genes and their genomic “flexibility.” The fact that cereals, despite the extensive conservation between grasses on gene content level as discussed earlier (Gill et al. 2004 for review), retained little of this conservancy for disease resistance genes to enable for rapid evolution (Leister et al. 1998) points toward such a natural selection. For example, in a comparative study by Du Preez et al. (2008)—wherein the largest group of known resistance proteins (i.e., nucleotide-binding site–leucine-rich repeat domain proteins, Jones and Takemoto 2004; Takken and Tameling 2009) associated with invader recognition during plant defense were studied across genera—supporting evidence for a divergence-before-duplication model of R-gene, evolution in triticale was found. These findings emphasized that despite the high level of conservation in the functional motif of resistance genes, there are also vast areas of variable regions within these genes that enable for mutational change and adaptative responses to specialist herbivory.

Plant adaptations to herbivory may be classified by the utilization of an additional biosynthetic pathway, an increased level of investment of defense resources, or by the degree of negative effect on the target organisms. In entomological context, three classes of insect responses to the plants’ defense suite can be formalized, namely antibiosis—where aphid biology is negatively affected by the host plant (Painter 1958), antixenosis—the plant is not a host of preference for D. noxia in terms of food, shelter, or reproduction (Painter 1958), and tolerance—the plant survives under levels of infestation that will kill or severely injure susceptible plants (Painter 1958). In an ecological context, tolerance rather than resistance may be advantageous if resources are relatively abundant (Fine et al. 2004) but a disadvantage when resources are limited. Chemical resistance characters may act as toxins, inhibitors of digestion, or as basic deterrents. One can also imagine that alteration in many plant features could knock out necessary stimuli for oviposition or feeding by certain specialized insects, and so inferring that compounds that act as deterrents could be more variable among plant taxa than toxins (Futuyma 2008; Futuyma and Agrawal 2009).

Diuraphis noxia feeding on the wheat plant for the purpose of growth and replication diminishes the performance of the plant and can even result in its death. In this parasitic relationship, once D. noxia’s presence on the wheat leaves is detected, a signal is generated to activate structural and biochemical defense mechanisms, in an attempt to deter the aphid (Van der Westhuizen et al. 1998; Botha et al. 2005, 2006, 2010; Boyko et al. 2006; Smith et al. 2005, 2010; Smith and Boyko 2007; Smith and Clement 2012; Murugan and Smith 2012). These defense mechanisms limit the chance of infestation, although present in most cultivars, sometimes fail because they are activated to slowly or not at all in susceptible plants (Van der Westhuizen et al. 1998). The key to an effective defensive system lies in the timely detection of the invading organism before it can settle and reproduce. When the host detects the presence of D. noxia infestation, a suite of “defensive syndromes” or pathways are activated either non-specifically (with general effectors) or specifically (with protein effectors) (Johal et al. 1995; Argrawal and Fishbein 2006). After effector recognition, the plant will purposefully commit suicide in the invaded cells or tissues utilizing a process known as the hypersensitive response. In the larger context, this genetically controlled program is part of a developmental process known as programmed cell death (PCD) and is an innate immunity system widely spread among all living organisms (even humans). PCD is activated by “flooding” these cells with toxic chemical compounds (i.e., H2O2) (Foyer and Noctor 2009) using elevated levels of reactive oxygen enzymes which results in cell death or necrosis (Van der Westhuizen et al. 1998; Botha et al. 2005, 2006, 2010; Smith et al. 2005; Smith and Boyko 2007; Boyko et al. 2006).

Even though traditional theory states that plants usually make use of two distinct biochemical pathways (jasmonic acid (JA)/ethylene (ET) associated—wounding during herbivory— and salicylic acid (SA) associated—based on the “type” of invader, that is, herbivores or pathogens to realize systemic acquired resistance and survival (Johal et al. 1995)), molecular evidence clearly points toward the utilization of less well-defined or intermediate pathways and/or networks when a plant is confronted with phloem-feeding pests like D. noxia (Botha et al. 2010; Van Eck et al. 2010; Van Eck 2011). In fact, wheat utilizes all available resources, even an intermediate between the JA and SA pathways (Fig. 3), in response to D. noxia feeding based on the type of resistance gene present and the D. noxia biotype (Botha et al. 2010; Smith et al. 2010). For example, in a study on a wheat variety containing the Dnx resistance gene, it was found that differential regulation of lipoxygenase (LOX) and Ω-3 fatty acid desaturase (FAD) genes were associated with JA signaling after infestation with the USA D. noxia biotype 1 (Boyko et al. 2006). However, the association of FAD expression with JA signaling could not be confirmed in another wheat variety containing the Dn4 resistance gene (Smith et al. 2010), and the authors ascribed this observation to the difference in host genetic background.

Fig. 3
figure 3

Simplified models of the different molecular pathways utilized by the host in response to D. noxia feeding. Dark colored text and solid lines indicate key pathways and/or activities, whereas light colored text and dotted lines show pathways with less involvement

Full size image

It was recently shown that aphid HAMP in wheat consist of 16 carbon fatty acids, JA-, abscisic acid (ABA)-, and ET-signaling genes, and direct aphid defense genes (Smith et al. 2010). In a follow-up study, Liu et al. (2011) demonstrated a dependency on the timely differential regulation (as early as 1–3 h post-infestation) of the JA-signaling genes [LOX and allene oxide cyclase (AOC)] and metabolism genes to ensure effective defense against D. noxia feeding in resistant varieties, suggesting unique JA signatures during host defense.

Wheat varieties use products of molecular pathways that result in antibiosis, antixenosis or tolerant defenses against D. noxia. Tolerant plants use passive resistance to cope with the removal of energy and nutrients, and the damaging effects of aphid-derived molecules on chlorophyll levels, rather than deterring the insect. Photosynthetic compensation is deployed after 6 h to cope with the aphid-associated stress. Photosynthetic compensation necessitates the upregulation of components of the photosystems (PS). Aphid feeding interferes with the electron transport chain from PSII to PSI, leading to photobleaching of chlorophyll. By upregulating components of the electron transport chain and the rapid replacement of damaged components of PSII, as well as increasing levels of enzymes involved in photoassimilation, these plants manage to retain active photosynthesis and prevent chlorosis from occurring. Tolerant cultivars generally do not employ oxidative bursts associated with hypersensitive responses (Botha et al. 2008). In antibiotic cultivars, defense is geared to the rapid activation of resistance mechanisms to obstruct D. noxia feeding and oviposition, even injuring the aphid. Antibiotic cultivars recognize D. noxia stylet penetration and respond by activating signaling cascades and a substantial influx of Ca2+ into the cytosol within 2–5 h after feeding begins (Botha et al. 2005, 2008, 2010; Smith et al. 2005, 2010). Results indicated that the induced signaling cascades lead to an oxidative burst and increased levels of SA. Levels of reactive oxygen enzymes are finely regulated by several systems involving iron homeostasis, RNA-binding genes, ABC transporters assisting in the movement of iron–sulfur clusters, and WRKY networks (Zaayman et al. 2009; Van Eck et al. 2010; Van Eck 2011). Deposition of callose and sealing off of sieve elements interferes with aphid feeding. The production of ROS such as H2O2 elicits programmed cell death. PCD is visible as localized necrotic lesions on the leaves and is directed at the prevention of a “feeding” attack. Pathogenesis-related gene expression is induced to provide long-term protection through SAR, and with enforced cell walls, the cultivar is more resistant to subsequent attack (Van der Westhuizen et al. 2002). And lastly, antixenosis is associated with the expression of volatile organic compounds (VOCs), and ethylene-mediated pathways might predominate in these plants (Zaayman et al. 2009; Botha et al. 2010; Smith et al. 2010). Thus, the antixenotic mechanism of defense constitutes a modification of the wounding response, with significantly more crosstalk between SA-mediated and ethylene-/JA-mediated pathways. These observations imply that rampant parallelism or convergence is possible. They also bear on the important question of whether plants’ defense profiles are optimized (as they might be if most lineages retain the same biosynthetic capacities).

Why an uneven enigmatic arms race?

Even though aphids seemingly hold the competitive edge above plants in their “arms race” because of their long evolutionary time, as well as their long-standing association with their respective endosymbionts, with the intervention of humans, plants remain competitive in the evolutionary arms race. As noted by Smith and Clement (2012) that through the use of advances in molecular technology, arthropod resistance genes can be “identified, tracked, and manipulated” and then introduced into arthropod-resistant crops for improved durable resistance. A combination of these improved resistant varieties and integrated pest management practices with comprehensive biological, chemical, and cultural control strategies may provide the crop species with the must-needed edge to stay ahead of this arms race.

Also, as noted, “coevolution” has several meanings and may refer to population-level processes of reciprocal adaptation of interacting species. These interactions may be relatively specific or pairwise, that is, between two species where each adapting to a characteristic of the other. However, the interaction may also be more diffuse and multispecific. In such an interaction, a species adapts to the properties of a set of interactors, and so has genetically correlated responses to several species (Hougen-Eitzman and Rausher 1994). Natural plant populations are often found to be extremely diverse in their resistance to invaders, a characteristic lacking in domesticated crop species. The geographic mosaic theory of coevolution depicts geographical differences in selection intensity to be inherently part of coevolutionary processes. Where some communities will experience reciprocal selection (coevolutionary hotspots), while other sites will have little or no coevolutionary activity (coldspots) producing a mosaic that continually change (Thompson 1999). Therefore, at a microevolutionary level, diversity in adaptations of plants to herbivores and vice versa could range from ongoing interactions between the products from resistance genes and their respective Avr gene products to competing biochemical pathways reflected in antagonistic lineages. This competition may require relatively short time lags between reciprocal evolutionary changes, to induce “the decoupled, sequential bursts of adaptation and diversifications” as portrayed in Ehrlich and Raven’s “escape-and-radiate” scenario (Ehrlich and Raven 1964).