AFLP-Based Genetic Linkage Map of the Colorado Potato Beetle Leptinotarsa decemlineata: Sex Chromosomes and a Pyrethroid-Resistance Candidate Gene
David J. Hawthorne


A genetic linkage map was constructed from an intraspecific cross of the Colorado potato beetle, Leptinotarsa decemlineata. This is an initial step toward mapping the loci that underlie important phenotypes associated with insect adaptation to an agroecosystem. The map was made with 172 AFLP and 10 anonymous codominant markers segregating among 74 backcross (BC1) individuals. Markers were mapped to 18 linkage groups and a subset of the markers with a mean intermarker distance of 11.1 cM is presented. A pyrethroid-resistance candidate gene, LdVssc1, was placed onto the map as well. The sex chromosome was identified by exploiting the XO nature of sex determination in this species using patterns of variation at LdVssc1 and the codominant markers.

UNDERSTANDING the genetic basis of pest adaptation contributes to our understanding of evolution in changing environments and to the management of pest adaptation in agricultural ecosystems. Many pest insects have undergone rapid evolution during the course of their adaptation to the agricultural ecosystem. The traits that evolve might include general features of pest biology such as colonization ability, rapid generation time, and high fecundity (Hawthorne and Via 1994), or they may be focused on particular attributes of their plant hosts and control measures such as insecticides (Roush and McKenzie 1987; Heinrichs and Rapusas 1990; Heckelet al. 1997). As with all adaptive evolution, this process requires selection of phenotypes and a genetic basis to the phenotypic variation under selection. Pest adaptations to a particular agroecosystem provide a unique opportunity to observe both the process and the genetic architecture of selection response under conditions of natural selection.

The Colorado potato beetle [Leptinotarsa decemlineata (Say), hereafter CPB] is an important pest of potato (Solanum tuberosum L.) in North America, Europe, and Asia. It is native to North America but new to potato, shifting from its ancestral hosts (a suite of Solanum spp.) in the mid-19th century (Casagrande 1987; Jacques 1988). The geographical range of CPB has expanded from central and northern Mexico and the southern plains of the United States to now include virtually all potato-growing regions of the world. Populations of this insect have evolved resistance to many synthetic insecticides, including representatives of all major classes of compounds deployed before the 1990s (Argentineet al. 1989; Roushet al. 1990). Because of its status as a threat to a food staple, much ecological, physiological, and toxicological work has already been done, especially regarding this insect’s interactions with environmental compounds (both plant and human produced), pathogens, and parasitoids. Genetic aspects of this insect were first studied by Tower (1906) in what may have been the first (albeit possibly discredited) studies of mutation in animals (Sokoloff 1966). Karyotypic analysis reveals 17 autosomes and an XO sex determination system in which males are XO and females XX (Hsiao and Hsiao 1983).

The linkage map presented here was constructed primarily with amplified fragment length polymorphisms (AFLPs) as the genetic markers. AFLP is a PCR-based method that offers an efficient and reproducible means of generating the genetic markers needed for linkage map construction (Voset al. 1995). AFLP screens many loci quickly and reveals large numbers of polymorphisms. No previous genetic information on the study organism is necessary before use of AFLP and, because it is a PCR-based marker, small amounts of template DNA are sufficient for analysis. These characteristics invite comparison to randomly amplified polymorphic DNAs (RAPDs), another easy-to-use marker (Williamset al. 1990). Although a number of excellent linkage maps have been constructed using RAPD polymorphisms (Grattapaglia and Sederoff 1994; Hunt and Page 1995; Bradshawet al. 1998; Beeman and Brown 1999), several researchers have reported difficulty in reproducing RAPD genotypes (Riedyet al. 1992; Penneret al. 1993). AFLP offers improved reproducibility compared to RAPD because of more stringent reaction conditions (Voset al. 1995). Both AFLP and RAPD are dominant markers; among the F2 offspring from a cross of different homozygotes (e.g., AA and aa) the dominant homozygote (AA) and heterozygote (Aa) are indistinguishable. The reduced information content of dominant markers is not a large liability in mapping studies that utilize a backcross or testcross design; many marker loci will have only two genotypes segregating among the backcross offspring: heterozygotes (Aa) and recessive homozygotes (aa).

Here I report the mapping of AFLP and sequence-tagged AFLP fragments to 18 different linkage groups in CPB. The development of this linkage map lays an important foundation for future genomics research on the CPB, providing a means for localizing genes of ecological importance and revealing the genetic architecture of response to natural selection. The immediate purposes of this work are (1) to estimate recombinational distance between markers for use in quantitative trait loci (QTL) mapping and population genetics, (2) to develop codominant anchor loci for future mapping, (3) to identify the sex chromosome and develop reliable markers on that chromosome, and (4) to place a candidate locus for resistance to pyrethroid insecticides (LdVssc) onto the linkage map.


Beetle sources and genetic crosses: This study describes a backcross analysis of CPB from two pest populations [Long Island, NY (LI) and Freeville, NY (FR)] that differ significantly in their susceptibility to a number of insecticides (Roushet al. 1990). Each population originated from a single egg mass of ∼30 individuals. Parental beetles, a LI male and a virgin FR female, were mated and their offspring reared to adulthood on greenhouse-grown potato plants in whole-plant sleeve cages. An F1 female was mated to a second LI male and the backcross offspring (BC1) were reared to pupation, also on caged potatoes. Late fourth instar larvae were placed into a covered pupation chamber (28 cm × 54 cm × 6.25 cm) containing 6 cm of moist potting soil into which larvae readily burrowed and pupated. Adults were collected as they emerged from the soil and were housed in filter paper-lined petri dishes for 1 day before they were weighed and then sexed (Pelletier 1990). A total of 74 backcross individuals (33 males, 46 females) were genotyped to construct the linkage map.

AFLP construction: AFLP was first described by Vos et al. (1995) as a means of amplifying a random array of restriction fragments. The methods of Vos et al. (1995) have been modified here. During the early development of AFLP for this insect it became clear that too many bands were produced using the standard six-base recognition restriction enzyme (six-cutter) + four-cutter strategy. To remedy this condition, AFLPs were constructed from fragments created by two six-cutters, PstI and EcoRI, both to reduce the fragment number and to bias restriction toward nonmethylated DNA. Frozen insects were homogenized and template DNA was extracted using a commercial protocol (QIAamp tissue kit, QIAGEN, Chatsworth, CA). DNA was eluted into 400 μl of the supplied buffer and stored at -20°. AFLP constructs were assembled in a single-tube reaction by mixing 2.0 μg of genomic DNA, New England Biolabs (Beverly, MA) #4 restriction enzyme buffer [20 mm tris-acetate, 10 mm magnesium acetate, 50 mm potassium acetate, 1 mm dithiothreitol (DTT)], 1.8 mm DTT, 2 mm ATP, 100 ng/μl bovine serum albumen, 20 units PstI, 20 units EcoRI, 6 units T4 DNA ligase, and 5 pm of each double-stranded adapter (Table 1). The tubes were incubated for 14 hr at 37°. During this reaction the restriction enzymes create “sticky ends” to which the adapters are ligated. Adapter ligation destroys the restriction site and prevents subsequent cleavage.

View this table:

The sequences of adapters, primers, and primer combinations

Amplification: A two-step amplification strategy was used, similar to that described by Vos et al. (1995). The first step (preamplification) is a PCR with core primers that are complementary to the adapter sequence only (Table 1). The second amplification uses longer selective primers, which have two or three overhanging nucleotides at the 3′ end (Table 1). The preamplification round was standard PCR [20 mm Tris-HCl (pH 8.4), 50 mm KCl, 1.5 mm MgCl2, 0.2 mm dNTPs, 1 unit Taq DNA polymerase] that included 2 μl of the AFLP construction as template and 5 pm of each core primer (Table 1). This reaction was cycled 20 times for 1 min at 94°, 1 min at 56°, and 1 min at 72°. Preamplification products were diluted 1:2 with 10 mm Tris (pH 8.0) and stored at -20°. The second step, selective amplification, was also standard PCR (same reagents and concentrations), which included 1 μl of the diluted preamplification product as template, and one each of the EcoRI and PstI selective primers (Table 1). The reaction profile was a touchdown-PCR in which the annealing temperature was high (65°) for the first round and then reduced 0.7° for each of the next 12 cycles. The denaturing and extension stages for each cycle were 94° for 10 sec and 72° for 90 sec, respectively. This ramping-down of the annealing temperature was followed by 25 cycles of 94° for 10 sec, 56° for 40 + 1 sec per cycle, and 72° for 90 sec. Following selective amplification, an equal volume of loading dye (98% formamide, 10 mm EDTA, 0.025% xylene cynol, 0.025% bromophenol blue) was added to each sample. After denaturing at 95° for 2.5 min, 1-3 μl were loaded on a 4.3% denaturing polyacrylamide gel (30 × 40 cm). Gels were run for 4 hr at a constant 60 W and then silver stained (Promega, Madison, WI). The amplified fragments were visualized on a light box, recorded as present or absent, and digitally archived by scanning the gel directly. Two individuals scored all gels, and a third person reviewing archived images resolved discrepancies.

Marker-discovery attributes: To evaluate the usefulness of AFLP for the CPB, the abundance of AFLP polymorphisms was recorded for a random sample of five AFLP reactions performed on the entire backcross family. The numbers of bands per gel, polymorphic bands, and segregating informative bands were recorded.

Genotyping: Linkage analysis was performed on the markers generated from 23 AFLP reactions (Table 1). Markers were included in the analysis if they displayed a segregation pattern consistent with the backcross of a heterozygous F1 female to a homozygous recessive “tester” male (Aa × aa). In practice, AFLP bands were scored as present or absent in the BC1s only if they were present in the LI or FR parent, absent in the other, present in the F1, and absent in the backcross tester parent (LI male). All backcross offspring (BC1) inherited one allele at each locus from the backcross parent and one from the FR or LI parent via the F1. Using the “A-H” scoring scheme in Mapmaker 3.0 (Landeret al. 1987), BC1 individuals with markers from the LI parent were scored as “A” and those with markers from the FR parent as “H.”

Segregation analysis: Mendelian inheritance of all markers was tested by chi-square, with significance levels corrected for multiple comparisons (α/173 markers) to give an experimentwise α = 0.05 (χ2 > 13.1). MAPMAKER 3.0 was used to infer linkage relationships and to determine the marker order within linkage groups. Linkage was determined at LOD ≥ 3.0 at a maximum (Kosambi) distance of 30 cM and marker order was determined at LOD ≥ 2.0. Only 1 of the 173 markers could not be assigned to a linkage group at LOD ≥ 3.0. A subset of the assigned markers, which were approximately evenly spaced and could be ordered unambiguously at LOD ≥ 2.0 (Figure 1), was selected.

Cloning of AFLP bands: Single-locus markers were developed from AFLP loci at known map positions to act as anchors to a particular linkage group. Bands were isolated directly from the silver-stained gels, reamplified to increase their abundance, and ligated into the 2.9-kb TA cloning vector (pCR vector; Invitrogen, San Diego) according to the manufacturer’s protocol (Choet al. 1996). Cloned fragments were sequenced and primers were designed to amplify the cloned fragment from genomic DNA. The resulting amplified fragments were evaluated for size polymorphism (using agarose or denaturing polyacrylamide gel electrophoresis) or sequence polymorphism using double-stranded conformational polymorphism (DSCP; Atkinson and Adams 1997). If markers segregated appropriately as codominant polymorphisms, they were placed on the linkage map using the “try” command in MAPMAKER.

Placement of the candidate gene LdVssc1: Using primers published by Lee et al. (1999), a 349-bp fragment of the voltage-sensitive sodium channel LdVssc1 gene, including the functionally significant L1014F mutation site, was amplified from all members of the mapping family. This locus is similar to the Drosophila sodium channel gene para, which has been shown to contribute to target-site insensitivity to pyrethroid insecticides in CPB and other pest insects (Dong and Scott 1994; Guerreroet al. 1997; Parket al. 1997). Fragments were evaluated by DSCP and genotypes of all individuals were determined.


AFLP marker-discovery attributes: The six + six-cutter strategy reduced the number of bands per gel sufficiently to resolve individual bands using +2 and +3 base overhanging primers. The five reactions evaluated for marker attributes had an average of 72.4 bands on each gel. Fifty percent of the bands were polymorphic within the backcross family, and 14.4% of the bands were segregating appropriately for use in the linkage analysis. Band sizes ranged from ∼200 to 1500 bp, with most in the 300 to 900-bp range. As expected, fragments are larger than those previously reported for another insect, the fall armyworm (McMichael and Prowell 1999), in which the six-cutter + four-cutter strategy was used. The 23 AFLP reactions generated 173 informative markers. The expected segregation ratio of each marker in the BC1 generation is 1:1 and none differed significantly from this ratio (all χ2 < 13.1).

Linkage map construction: MAPMAKER inferred 18 linkage groups from the backcross data (Figure 1). Because the haploid chromosome number of CPB is 18 (Hsiao and Hsiao 1983), our linkage groups appear to correspond to chromosomes. Of the 173 AFLP markers used to make the map, positions of 86 well-spaced AFLPs, 10 codominant markers derived from AFLPs, and the LdVssc1 fragment are presented here. The map covers 1032 cM with a mean (±SD) intermarker distance of 11.1 cM (±6.72).

Cloned AFLP fragments: I chose 10 AFLP markers that segregated in the backcross population and occurred on different linkage groups in preliminary analyses to be cloned. Table 2 lists the primer sequences and fragment sizes for the cloned markers. Map locations of these single-locus PCR markers are indicated on Figure 1. All single-locus markers were found on the same linkage groups as their progenitor AFLP marker (Figure 1), with differences in recombination distance between loci most likely due to genotyping error in the AFLP analysis. Sequences of single-locus markers have been submitted to GenBank (accession nos. AYO28388-AYO28397).

Placement of candidate gene LdVssc1: All individuals in the mapping population were genotyped using the fragment of LdVssc1 described by Lee et al. (1999). The “try” command in MAPMAKER was used to place this fragment on linkage group 11 (LOD = 8.21), and its map-order position on linkage group 11 (Figure 1) was supported by a LOD score of 4.11.

Determination of the X chromosome: The CPB, like many insects, has an XO system of sex determination in which females are diploid (XX) and males are haploid (XO). This systematic difference in ploidy was exploited to distinguish the X chromosome from the autosomes, which are all diploid for males and females. Previous work (Argentineet al. 1989; Heimet al. 1992) suggested that pyrethroid resistance is sex linked, offering a starting point for our search (as the candidate gene for pyrethroid resistance, LdVssc1, mapped to linkage group 11). The genotypes at LdVssc1 of the F1 and BC parent were Aa and aa, respectively. If this locus is on the X chromosome there will be four genotypes of BC1 offspring in which the males are A_ or a_ and females Aa or aa. Using DSCP to visualize the amplified fragments, a single band may be due to homozygosity or hemizygosity. Therefore, the “aa” and the “a” genotypes are indistinguishable and cannot be used to distinguish XX individuals from XO. However, all “Aa” individuals (21/21) were female and all “A” individuals (17/17) were male, fitting the pattern expected if LdVssc1 were on the X chromosome. A second marker, the codominant SL-1010, also maps to linkage group 11. Using DSCP again, 43 of 44 (98%) BC1 females were heterozygous (diploid), while 29 of 30 (97%) males were hemizygous, with all hemizygous males receiving their allele from the maternal and none from the paternal parent. None of the other nine single-locus codominant markers mapping to different linkage groups displayed sex-related differences in banding pattern. As an additional check, 20 field-collected beetles were sexed and genotyped at the SL-1010 locus. Whereas 9 of 10 females were demonstrably diploid (two bands), all 10 males had a single band, suggesting haploidy. These results strongly indicate that linkage group 11 corresponds to the X chromosome.

Figure 1.

—Linkage map of the CPB genome. AFLP marker loci are labeled as described in Table 1; single-locus markers are named with the approximate size of the progenitor AFLP marker and prefaced by “SL.” Two markers indicated with an asterisk are ordered at LOD ≥ 1.6; all others are ordered at LOD ≥ 2.0.


By considering the evolution of ecologically important traits in pest insects much can be learned about the genetic basis of phenotypic evolution, especially for populations faced with strong selection. Remarkably, the genetic basis of adaptation in natural populations is largely unknown (Orr 1998). The use of genetic linkage maps can help identify the loci contributing to adaptive changes in populations. Maps with a density of markers around 20-30 cM are sufficient for detecting the presence of QTL with increased marker density allowing more precise positioning of the QTL (Darvasi and Soller 1994; Lynch and Walsh 1998). The 11-cM map presented here will allow efficient QTL mapping in this insect. One important use of genetic linkage maps is comparison of positions of mapped candidate loci with QTL. Such comparison will streamline efforts to associate ecologically important genetic variation with known genes. This study confirms that a candidate gene, LdVssc1, previously shown to contribute a large degree of resistance to pyrethroid insecticides (Leeet al. 1999), is located on the X chromosome (linkage group 11) in the CPB. The resistant phenotype will therefore be expressed in all males carrying the resistance-conferring allele, even if it is recessive, providing a boost to the rate of increase of a favored allele especially when that allele is rare. Now, we can ask whether QTL associated with resistance to pyrethroid insecticides are found at or near this locus on linkage group 11 and whether genetically correlated resistances to other insecticides are due to pleiotropic effects of that locus.

View this table:

Names, primer sequences, and characteristics of single-locus markers

The CPB linkage map will also contribute to population genetic goals. For example, knowing the recombinational distance between a locus that has undergone strong selection (such as LdVssc1) and other markers allows us to measure the effect of that selection on nucleotide variation at linked loci (e.g., Yanet al. 1998). This will further efforts to develop relationships between genetic diversity, strength of selection, and genetic distance from the locus under selection in contemporary populations.

AFLP seems especially well suited for linkage map construction using intraspecific crosses. In this study, variation was so abundant that marker screening was unnecessary. The mean number of AFLP markers identified per gel (7.4) was much higher than that observed using RAPDs on Apis melifera (Hunt and Page 1995; 2.8 bands per gel) or on Tribolium castaneum (Beeman and Brown 1999; 1.5 bands per gel). However, because CPB populations may have inherently higher levels of genetic polymorphism than either Apis or Tribolium, these differences may not be due solely to the markers.

Use of AFLP for mapping represents a fundamentally different approach to identifying and mapping loci than that typically used with RFLP or microsatellites. Construction of linkage maps with AFLP allows us to discover and localize markers of interest and then invest in converting them into codominant markers. In contrast, RFLP or microsatellite approaches require us to invest in marker development first and then determine if they are useful. Following this strategy here, I have cloned and sequenced several AFLP fragments, converting them to single-locus codominant markers. These features make AFLP potentially useful in many types of population studies and may increase the accessibility of genetic information for genetically unstudied species.


Andrea Badgley, Maria Loosen, Stasia Skillman, and Mike Jackson helped in the lab, and Dale Moyer and Ward Tingey supplied insects. Fred Gould, George Kennedy, and Susan McCouch provided lots of early inspiration. Jeffrey Feder, John McKenzie, Yvonne Parsons, Don Price, and Kerry Shaw made many helpful comments. I especially thank Richard Harrison who generously provided research space and guidance during the initial phases of this work. This work was first supported by the Center for Integrated Pest Management and then by grant no. 58-1275-6-034 from the U.S. Department of Agriculture.


  • Communicating editor: G. A. Churchill

  • Received March 7, 2000.
  • Accepted March 2, 2001.


View Abstract