Identifying Resistance to the Colorado Potato Beetle (Leptinotarsa decemlineata Say) in Potato Germplasm: Review Update

Developing plant germplasm that contains genetic resistance to insect pests is a valuable component of integrated pest management programs. In the last several decades, numerous attempts have been made to identify genetic sources of resistance to Colorado potato beetle Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae). This review focuses on compiling information regarding general L. decemlineata biology, ecology, and management focusing on discussing biochemical and morphological potato (Solanum tuberosum L.) plant traits that might be responsible for providing resistance; the review ends discussing past efforts to identify genetic material and highlights promising new strategies that may improve the efficiency of evaluation and selection of resistant material. Measurement strategies, that begin with field screening of segregating populations or wild germplasm to narrow research focus can be useful. Identifying particularly resistant or susceptible germplasm, will help researchers focus on studying the mechanisms of resistance in much greater detail which will help the development of long-term sustainable management program.


INTRODUCTION
Hundreds of crop varieties are released every year to improve crop productivity, enhance nutritional value, and expand consumer choice (Everson and Gollin, 2002); crops are also improved to repel, prevent, block, or eliminate pests (Mohammed et al., 2000;Rondon et al., 2009). Clearly, crop improvement is one of the drivers of crop production and extensive studies have evaluated the contributions of plant breeding to production enhancement in association with marketing and utilization (Douches et al., 1996). However, when we manage pests including insects, diseases, or nematodes, the limitations of biological and other control tactics have pushed producers to rely on the use of pesticides as we will discuss below (Mitkowski and Abawi, 2003;Ramirez et al., 2009;Maharijaya and Vosman, 2015).
In commercial potato, Solanum tuberosum L., the Colorado potato beetle, Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae), is well-known for rapidly evolving resistance to insecticides (Alyokhin et al., 2008;Schoville et al., 2018), which reinforces its status as one of the most important potato pests (Weber, 2003). Over 300 documented cases of insecticide resistance are listed in the literature coming from Asia, Europe, and North America (Whalon and Mota-Sanchez, 2017;Brevik et al., 2018a,b). Interestingly enough, populations of L. decemlineata do not respond equivalently to pesticides; some researchers argue that it may be due to evolutionary differences in how L. decemlineata interacts with potatoes within the landscape (Whitaker, 1994;Wierenga and Hollingworth, 1994;Crossley et al., 2019a,b). Others suggest a moderate variation at the nuclear loci level (Hawthorne, 2001;Izzo et al., 2018). Remarkably, in Idaho, Oregon, and Washington which together produce close to 56% of United States fresh and frozen potatoes in the market, L. decemlineata populations have largely remained susceptible to insecticides (Haegele and Wakeland, 1932;Johnston and Sandvol, 1986;Olson et al., 2000;Alyokhin et al., 2015;Crossley et al., 2018;Dively et al., 2020). According to Lynch and Walsh (1998), populations with lower genetic variance are expected to adapt more slowly, and L. decemlineata in the Northwestern United States exhibits lower genetic diversity than elsewhere which may explain its susceptibility to pesticides. In comparison, populations in other areas like east of the Rocky Mountains, L. decemlineata has historically developed pesticide resistance (Grafius, 1995;Grafius and Douches, 2008). Hence, the need to add novel tools in the pest management toolbox to help potato producers manage this pest problem remains and the availability of L. decemlineata potato resistant varieties could complement existing management practices.
In the US, L. decemlineata was first associated with Solanaceous plants around 1859, when it was reported on S. rostratum in Omaha, Nebraska, near the Iowa border (Casagrande, 1985). By 1859, L. decemlineata was found on S. tuberosum (Walsh, 1866), and by 1874, L. decemlineata spread rapidly eastward reaching the Atlantic coast (Tower, 1906;Hsiao, 1978;Logan et al., 1987;Jacques, 1988;Alyokhin, 2009;Zhao et al., 2013;and Izzo et al., 2018). Member of the genus Leptinotarsa have been long known to feed on S. rostratum in central Mexico, suggesting Mexico as the center of distribution of L. decemlineata. Historically, as the Solanaceous weed geographical range expanded northward, so did L. decemlineata distribution (Table 1). Worldwide, by the end of the 20th century, L. decemlineata was well-established in Europe, Asia Minor, Iran, Central Asia including China (Jolivet, 1991;Weber, 2003;Alyokhin, 2009). In Europe, L. decemlineata became established near Bordeaux, France in 1922, and spread throughout the French potato production regions, although it went undetected until 1935 (Hurst, 1975). Hurst (1975) reported three invasions (1876,1901,1914) in Great Britain that were "aggressively" (chemically) controlled. Crossley et al. (2017) reported that currently in Europe there are two clades: a western clade (French, Spanish, and Italian) and an eastern one (Poland, Estonia, Finland, and Russia). This information suggests multiple arrivals of L. decemlineata in Europe but could also be explained by adaptation to regional environment/hosts or genetic drift within isolated populations. Moving to Asia, L. decemlineata populations reached western China by 1993 and now it can be found in several Chinese potato regions (Zhang et al., 2013).
Recently, Izzo et al. (2018) used mitochondrial DNA and nuclear loci to examine the origin of L. decemlineata lineages. Authors suggested a genetically heterogenous L. decemlineata population which may contribute with contemporary populations. Schoville et al. (2018) used L. decemlineata as a model species for agricultural pest genomics that provides a better understanding of the insect and its host information that can be used for future genetic and evolutionary studies.

LIFE HISTORY OF LEPTINOTARSA DECEMLINEATA
Adults and larvae of L. decemlineata are voracious defoliators of S. tuberosum. Ferro et al. (1985) report that approximately 40 and 10 cm 2 of potato leaves are consumed by larvae and adults during their lifetime, respectively. Once leaf tissue is gone, L. decemlineata begins feeding on stems and tubers (Weber and Ferro, 1993), before moving into the soil to overwinter.

CURRENT MANAGEMENT PRACTICES
Core management of L. decemlineata focuses on chemical, cultural, and biological control.
Crop rotation (Lashomb and Ng, 1984;Weisz et al., 1994;Speese Iii and Sterrett, 1998;Huseth et al., 2012), the use of plastic-lined trenches (Boiteau and Vernon, 2001), straw mulch (Stoner et al., 1996;Stoner, 1997), trap cropping (Hunt and Whitfield, 1996;Hoy et al., 2000), thermal control (Rifai et al., 2004), and electromagnetic control (Colpitts et al., 1992) are a few examples of cultural, mechanical and physical control options. Interestingly, combining the dispersal behavior of L. decemlineata, which despite their ability to fly mainly disperse by walking, and the rotation of potato fields with non-host crops can reduce adult abundance (Follett et al., 1996;Hough-Goldstein and Whalen, 1996;. More recently, Crossley et al. (2017Crossley et al. ( , 2019a discussed the effect of land cover composition influencing the efficacy of crop rotation such as higher wheat land in the Northwest and forest field edges in the Central Sands of the US. These studies suggest the importance of land coverage in L. decemlineata movement and distribution and further studies are currently underway in the Northwest. Some field studies suggest that natural enemies have minimal effect Koss et al., 2004Koss et al., , 2005. However, predators such as the Carabidae Pterostichus melanarius Illiger (Alvarez et al., 2013, Rondon et al., 2013 and the Chrysopidae Chrysoperla carnea Stephens (Sablon et al., 2013) were found feeding on L. decemlineata. Two Tachinidae parasitic flies, Myiopharus aberrans L. and M. doryphorae Riley (Figure 2) were found to parasitize L. decemlineata with relative success (Kelleher, 1966;Tamaki et al., 1982Tamaki et al., , 1983Gollands et al., 1991;Lopez et al., 1993Lopez et al., , 1997. Several species of Bacillus including B. pumilus L., B. cereus Frankland and Frankland, B. megaterium de Bary (Ertürk et al., 2008), and B. thuringiensis Berliner (Walker et al., 2003;Whalon and Wingerd, 2003;Gassmann et al., 2009;Wraight and Ramos, 2015) were found to cause L. decemlineata mortality. Although the real impact of natural enemies controlling L. decemlineata is unknown, studies by Snyder and Clevenger (2004) and Lynch (2013) exploring the consumption by two generalist predators in potatoes using molecular gut content analysis and behavioral studies could provide new insights into host selection and role of microbiota into L. decemlineata fitness and sucess. For additional information about natural enemies see Table 2. HOST SELECTION BY L. DECEMLINEATA: THE NEED FOR "WILD RELATIVES" In herbivores, chemical signals have an important role in host plant selection (Fürstenberg-Hägg et al., 2013;Sablon et al., 2013;Wen et al., 2019). Volatiles like trans-2-hexen-1-ol-hexanol, cis 3-hexen-1-ol, trans-2 hexenal, and linalool, methyl salicylate, and z-3-hexenyl acetate have been reported as key cues for L. decemlineata (Visser et al., 1979;Dickens, 2000Dickens, , 2002Martel et al., 2005). Moreover, damaged leaves potentially produced and releases additional chemicals that may attract L. decemlineata (Boiteau et al., 2003). Stimulants, or essential dietary components produced by potato plants, released by potato plants such as sterols (e.g., cholesterol, b-sitosterol, stigmasterol) and sucrose, melezitose, glucose and fructose, amino acids, phospholipids, and chlorogenic acid (Hsiao, 1969) act as feeding stimulants (Szafranek et al., 2008); conversely, flowers of tansy (Tanacetum vulgare L.) which contain high levels of camphor and umbellulone, act as feeding deterrents (Yencho et al., , 1996(Yencho et al., , 2000Pelletier and Dutheil, 2006;Maharijaya and Vosman, 2015). Based on this information, manipulation of chemical cues which may affect "normal" behavioral choices of L. decemlineata potentially could be deployed via plant breeding to regulate feeding behavior and manipulate their control. A summary of current efforts will be discussed below.

PLANT BREEDING
Plant breeding is a powerful tool that can contribute to pest management. The notion of utilizing variation in L. decemlineata host preference between potato varieties (Torka, 1950;Horton et al., 1997;Metspalu et al., 2000) for breeding has been considered since the late 19th century (Saunders and Reed, 1871;Chavez et al., 1988). However, the potential to improve insect resistance by breeding within domesticated potato alone is limited due to potato narrows genetic base (Hardigan et al., 2017;Jansky and Spooner, 2018). Fortunately, the germplasm pool of approximately 107 Solanum spp. related to potato comprises one of the deepest and most accessible sources of pest resistant alleles among all major crops (Spooner and Bamberg, 1994;Jansky et al., 2013;Spooner et al., 2014;Bethke et al., 2017) (Table 3). Systematic field screens aimed at identifying crop wild relatives that exhibit L. decemlineata resistance have been performed on large scales by several investigators since the early 1930s (Torka, 1950), with the efforts of Carter (1987), Flanders et al. (1992), and Jansky et al. (2009) all making major contributions to our knowledge on this topic at a species and accession level (Pelletier et al., 2011).
Other glycoalkaloid compounds are also known to play prominent roles in plant defense against L. decemlineata, but less is understood regarding inheritance of pathway components at a molecular genetics level. Dehydrocommersonine and solanidenol-chacotriose concentration has been demonstrated to play opposing roles in L. decemlineata resistance when inherited from S. oplocense (Tai et al., 2015;Paudel et al., 2019). Selective genotyping of a S. oplocense × S. tuberosum F1 population have mapped a QTL associated with production of these molecules to a location on chromosome 1 (Paudel et al., 2019) also identified in a S. berthaultii × S. tuberosum F1 population by Yencho et al. (1998). Correlative evidence supporting association between L. decemlineata resistance of wild species and other glycoalkaloid molecules found in potato is intriguing but less direct in so far as proven causation.
Complimentary studies aimed at characterizing the inheritance of trichome related traits and L. decemlineata resistance in reciprocal S. berthaultii × S. tuberosum hybrid backcross populations indicate that traits exhibits some shared and some unique features to their genetic architecture (Bonierbale et al., 1994;Yencho et al., 1996). Inheritance of insect resistance is clearly a multi-locus trait that exhibit differences in penetrance based upon how trait values are skewed differentially between backcross parent (Bonierbale et al., 1994;Yencho et al., 1996). The number of type A trichomes is largely controlled by the inheritance of S. berthaultii alleles on chromosome 6 (dominant) and 10 (recessive), which combine to describe up to 63% of variation in a S. berthaultii backcross population (Bonierbale et al., 1994). PPO concentration is also a multi-locus trait with QTL on chromosomes 2, 5, 8, and 11 potentially contributed to enzyme abundance (Bonierbale et al., 1994). Inheritance of type B trichomes is clearly recessive as type B trichomes are not observed in S. tuberosum backcross populations derived from S. berthaultii × S. tuberosum hybrids (Bonierbale et al., 1994). The presence of sucrose ester droplets and abundance of type B trichomes are controlled by a major QTL located on chromosome 5 and additional QTL located on chromosome 1, 2, and 4 (Bonierbale et al., 1994). These plant characteristics seem to directly relate with L. decemlineata feeding and reproductive presence in backcross populations, which are influenced by no fewer than four QTL located on chromosomes 1, 5, 8, and 10 (Yencho et al., 1996). When the results of both Bonierbale et al. (1994) and Yencho et al. (1996) were overlaid, L. decemlineata resistance as measured by Yencho et al. (1996) exhibited QTL located on 1, 5, 10 in a S. berthaultii backcross population whereas QTL associated with type A trichome characteristics (Bonierbale et al., 1994) were associated with QTL 1, 2, 4, 5, 6, 10, and 11 in the same population. Whereas, in a S. tuberosum backcross population fewer QTL were identified and the only major overlapping QTL was associated with PPO activity on chromosome 8 (Yencho et al., 1996). Largely due to the thoroughness of these investigations (Bonierbale et al., 1994;Yencho et al., 1996), it is suspected that chemical resistance also may play a role due to segregation of glycolalkaloid content associated with the S. berthaultii background (Yencho et al., 1998).
Although great strides have been made toward the identification, introgression, and genetic mapping of L. decemlineata resistance mechanisms into breeding lines, there is still much work to be done. Currently no L. decemlineata resistant cultivars have been widely accepted and planted on large acreage (Grafius and Douches, 2008). Linkage mapping is an indispensable tool for mapping quantitative traits, but other complimentary methods including genome-wide association studies (Sharma et al., 2018;Yuan et al., 2019) or quantitative evolutionary genetic methods (Hardigan et al., 2017;Li et al., 2018) can have a great impact identifying major allelic variants within populations/species groups. Comparative genomics with other Solanum spp. has long been perceived as a method to leverage our knowledge of biochemical genes across species (Cárdenas et al., 2015) and has been used to identify shared components responsible for glycoalkaloid biosynthesis (Itkin et al., 2011;Cárdenas et al., 2016).
Little is understood about the mechanisms of resistance present in other species outside of S. chacoense and S. berthaultii or the genetics underlying these traits. This is particularly apparent for germplasm from the endosperm balance number 1 clade in potato (Carter, 1987;Flanders et al., 1992;Jansky et al., 2009;Pelletier et al., 2011). Consistently high levels of resistance in field screens but the mechanisms responsible of this quality are less tractable to breeders focused on cultivated S. tuberosum due to crossing barriers. Bridge crosses (Yermishin et al., 2014) and somatic hybridization (Jansky et al., 1999) are tools that researchers have used to introduce these traits into backgrounds compatible with cultivated potato (Jansky, 2006).
Transgenic potatoes thru genetic engineering has focus on pest resistance (Michaud et al., 1993;Alyokhin and Ferro, 1999;Grafius and Douches, 2008;Šmid et al., 2013). In 2013, Cooper et al. (2004) combined genetic engineering and traditional breeding testing feeding, biomass accumulation, and mortality on three populations of L. decemlineata including a Bt cry 3A-selected population which apparently conferred elevated resistance in potatoes. Mi et al. (2015) also reported excellent results expressing cry3A genes. However, L. decemlineata adapts quickly (Zhu-Salzman and Zeng, 2015). Jin et al. (2015) developed large-scale tests of natural refuge for delaying cotton bollworm (Helicoverpa armigera Hüber) resistance to transgenic Bt crops that could be used for L. decemlineata. Li et al. (2017) also studied the function and effectiveness of natural refuge in Insecticide Resistance Management strategies for Bt crops but more studies are still needed. Ma et al. (2020) reviewed the benefits of RNA interference (RNAi) tested for L. decemlineata where target genes, efficiency, and factors affecting RNAi efficiency against this pest were discussed. Maharijaya and Vosman (2015) suggest that breeding potato for resistance against L. decemlineata will benefit from information obtained using new, higher throughput screening methods. This has proven true in row crops, particularly as it relates to quantitative genetics and marker assisted selection. Improvement of crops using quantitative genetics requires identifying biological association between the plant phenotype of interest (insect resistance) and genetic or biochemical markers that can be used to select for that trait (Bernardo, 2019;Paudel et al., 2019).

CURRENT BREEDING EFFORTS: NORTHWEST CASE STUDY
Oftentimes the number of measurements required for genetic mapping, particularly in field studies, substantially outnumber the resources that can be allocated to perform this task by any single research group. In such cases, collaborative efforts between domain experts (entomology, breeding, remote sensing, analytical chemistry etc.) can be hugely beneficial. Small scale field studies can help us assess our ability to evaluate insect preference. Such information is critical for understanding both the insect pressure at a given trial location and how the aggressiveness of foliar consumption varies between field seasons.
Human index scoring of insect mediated defoliation is a relatively quick and accurate measurement, however as the clone number within a trial and number of field sites increases, our ability to accurately record this data becomes limiting. Data collection using multispectral sensors fixed to sUAS have been proposed as a scalable method to estimate insect damage in numerous crops (Puig et al., 2015;Mahlein, 2016;Vanegas et al., 2018) including potato (Hunt et al., 2016;Hunt and Rondon, 2017;Théau et al., 2020). When combined with ground truth measurements, multispectral data collected using sUAS can provide high-resolution, quantitative data on plant performance on larger scales and at greater frequency than could possibly achieved by human observation alone (White et al., 2012). This sort of standardized measurement is particularly attractive for multi-location field trials where scoring of insect mediated defoliation must be measured by members of different research groups and with high spatiotemporal resolution. Unfortunately, spectral measurements can be influenced by many confounding factors including, plant genotype, plant developmental stage, management practices within the trial, and environmental variables across time (White et al., 2012;Hunt et al., 2016;Khot et al., 2016;Chivasa et al., 2020;Théau et al., 2020). So, more research is needed to ascertain how reliable these measurements are at measuring insect damage specifically, across locations and field seasons.
We believe that multiscale measurement strategies, that begins with field screening of segregating populations or wild germplasm to narrow research focus can be useful and are an efficient use of resources. Identifying particularly resistant or susceptible germplasm, will help researchers focus on studying the mechanisms of resistance in much greater detail though application of some of the more difficult, expensive, and time demanding assays described in this article.

AUTHOR CONTRIBUTIONS
SR and MF planned the idea of the review and wrote entire manuscript. SR designed the framework of the review. AT contributed with the writing and editing and providing information for Table 3