No severe genetic bottleneck in a rapidly range-expanding bumblebee pollinator

Genetic bottlenecks can limit the success of populations colonizing new ranges. However, successful colonizations can occur despite bottlenecks, a phenomenon known as the genetic paradox of invasion. Eusocial Hymenoptera such as bumblebees (Bombus spp.) should be particularly vulnerable to genetic bottlenecks, since homozygosity at the sex-determining locus leads to costly diploid male production (DMP). The Tree Bumblebee (Bombus hypnorum) has rapidly colonized the UK since 2001 and has been highlighted as exemplifying the genetic paradox of invasion. Using microsatellite genotyping, combined with the first genetic estimates of DMP in UK B. hypnorum, we tested two alternative genetic hypotheses (‘bottleneck’ and ‘gene flow’ hypotheses) for B. hypnorum's colonization of the UK. We found that the UK population has not undergone a recent severe genetic bottleneck and exhibits levels of genetic diversity falling between those of widespread and range-restricted Bombus species. Diploid males occurred in 15.4% of reared colonies, leading to an estimate of 21.5 alleles at the sex-determining locus. Overall, the findings show that this population is not bottlenecked, instead suggesting that it is experiencing continued gene flow from the continental European source population with only moderate loss of genetic diversity, and does not exemplify the genetic paradox of invasion.


(a) Genetic diversity and bottleneck analysis
later batches of eggs laid by the queen (Table S5). Hence, the two classes of colony were 165 exclusive in terms of timing of male production (males in first brood vs. no males in first 166 brood), but five of the first-brood male producing colonies went on to produce late males as 167 well (i.e. males were produced in both the first and subsequent broods). 168 For each male-producing colony, up to the first 10 adult males to eclose were selected for 169 genotyping. However, for colonies producing first-brood males and more than 10 males 170 across their entire colony lifespan (colonies 120, 138 and 159), or producing only late males 171 but producing at least one diploid male in the initial 10 adult males selected for genotyping 172 (colony 83), the sample size was increased, with the first 24 males to eclose being selected 173 for genotyping (Tables S3, S4, S5). The sampling date of each individual sampled male was 174 noted, and as all sampled males were sampled shortly after their eclosion (i.e. as callows), 175 this allowed the sequence of eclosion for all genotyped males within a given colony to be 176 known (Tables S4, S5). Using these methods, a total of 84 males were sampled from the first-177 brood male producing colonies, comprising 25 first-brood males (defined as those males that 178 eclosed within one week of first worker eclosion or were produced instead of workers) and 179 59 late males (defined as those males that eclosed later than one week after first worker 180 eclosion). Further, a total of 148 males (all of which were late males) were sampled from late 181 male producing colonies. Hence, a total of 232 adult males from 28 field-collected queen-182 reared colonies were selected for genotyping, comprising 25 first-brood males (i.e. all 183 available first-brood males produced by first-brood male producing colonies) and 207 late 184 males (Table S3). All sampled males were kept at -20°C until DNA extraction. 185 For all males (n = 380 pupal males from field-collected nests plus 232 adult males from 186 colonies reared from field-collected queens = 612 males in total), DNA was extracted from 187 the thoracic tissue of each male (pupal or adult) and PCRs were performed using the methods 188 described for workers above [2]. Similarly, all male samples were genotyped at the same 14 189 microsatellite loci as those used for the workers. Each PCR plate included (1) a negative 190 control, comprising all reagents and primers but no template DNA, and (2) a positive control 191 comprising a single haploid male or a single diploid queen with known genotypes. Given that 192 some colony queens produced only a single male, alleles were accepted if they appeared in 193 one or more males, so that rare alleles in the population were not missed. The negative 194 controls did not show any peaks corresponding to the amplified alleles. Marker BTMS0132 195 was later found to be monomorphic across all typed males and was therefore dropped from Males were considered diploid if they were phenotypic males (Figures S1, S2) that were 201 heterozygous at two or more of the microsatellite loci. All males in colonies (both field-202 collected and those reared from field-collected queens) in which diploid males were detected 203 (by this criterion) during the first round of genotyping (field-collected colonies: n colonies = 2, 204 n males = 48; colonies reared from field-collected queens: n colonies = 6, n males = 70) underwent 205 re-extraction of DNA and re-genotyping, to ensure that diploidy was not called on the basis 206 of contamination. In two colonies, two males (56MP5 and 135M1), representing 0.33% of all 207 genotyped males, appeared diploid during the first round of genotyping and haploid during 208 the second round of genotyping. Therefore, these males underwent DNA extraction and 209 genotyping for a third time, allowing them to be assigned as definitively haploid or diploid.

210
Both males appeared homozygous across all loci in the second and third rounds of 211 genotyping and were therefore classed as haploid. To ensure that haploidy of males sampled 212 from first-brood male producing colonies was not called on the basis of large allele dropout, 213 whereby alleles with longer sequence length are missed, all sampled males that appeared as 214 haploid from the first-brood male producing colonies (n colonies = 8, n first-brood males = 21, n late These two regenotyping procedures, accounting for 25.8% of all sampled males, were also 217 used to calculate locus-specific allele error rates from mistyping for the male genotyping. On 218 this basis, the per-locus mean (range) allelic error rate was estimated to be 1.17% (0.00-219 7.25%) (Table S6). Allelic richness and allele frequency were calculated using Cervus v3.0.7.

220
Given that any diploid males are expected to show similar levels of heterozygosity to workers 221 (i.e., in the present case, having at least three heterozygous loci, since the minimum number 222 of these loci at which workers from the Norwich population [2] were heterozygous was 223 three), the initial scoring of two heterozygous loci for diploidy may not have been 224 conservative enough. Hence, to ensure that the number of loci at which males had to be 225 heterozygous to be considered diploid did not affect our results, the diploid male 226 classification was re-run. Here, two further criteria were introduced for diploid male 227 classification, with males classified as diploid if they were heterozygous at: i) one or more 228 loci (less conservative); or ii) three or more loci (more conservative).

229
Finally, the phenotypes of all males genetically confirmed as diploid (as described above) 230 were double-checked (by inspection of the original individual pupal and adult males sampled) 231 to confirm that they were not workers that had been mistakenly phenotyped as males (Figures 232 S1, S2). Double-checking of phenotypes confirmed all individuals classified as diploid males 233 by the above procedures were males, and not misidentified workers.

234
In summary, for a male to be assigned as diploid in the final data set, it was required to be 235 scored as heterozygous across at least two microsatellite loci during at least two independent 236 rounds of DNA extraction and genotyping, and also to have undergone double-checking to 237 confirm its male phenotype.

249
The frequency of colonies exhibiting DMP was estimated from the diploid male data in three 250 ways. The first used the frequency of colonies with DMP in the field-collected colonies 251 producing diploid offspring. The second used the frequency of colonies with DMP in the 252 colonies reared from field-collected queens that produced diploid offspring. The third 253 allowed for potential sampling error in determining DMP in the colonies reared from field-254 collected queens that produced diploid offspring, as some of these colonies produced only 255 small numbers of males for genotyping (Figure 2b,2c). This third method included, from the 256 colonies reared from field-collected queens that produced diploid offspring, only those 257 colonies in which the sum of the numbers of workers and/or males produced was three or 258 more. This was because the five colonies found to have produced diploid males (Figure 2b, 259 2c) all produced at least one diploid male among their first three males (Tables S4, S5).

260
Hence, if a colony were producing diploid males, it would have been expected to produce at 261 least one diploid male in its first three diploid offspring (including those for which the 262 fertilization event in fact led to their being sex-determined as workers), or in the first three 263 male offspring that were then selected for genotyping.

264
(iii) Estimation of allelic diversity at the sex-determining locus 265 From Adams et al. [20], if N is the number of alleles at the sex-determining locus in a system colony is headed by a single queen mated once [19,21] Let colonies headed by a queen mating 1, 2, or 3 times occur in the proportions X, Y, and Z, 283 respectively, and let the probability that each of these classes of colony contains a queen with 288 Therefore: Rearranged, this yields: producing colonies as outlined in the previous section: (1) from the field-collected colonies;

299
(2) from the colonies reared from field-collected queens; and (3) from the colonies reared 300 from field-collected queens after correcting for sampling error in the sampling of male 301 offspring for genotyping.

302
As the estimates of X, Y, and Z found in the study population by (mean queen mating frequency = 2.5), i.e. assuming either X = 1, Y = 0, Z = 0, or X = 0, Y = 308 0.5, Z = 0.5, respectively. For this, Equation S4 was used with D equal to the value estimated from the colonies reared from field-collected queens after correcting for sampling error, as 310 this was assumed to be the most accurate estimate of D.

312
Diploid male production and allelic diversity at the sex-determining locus 313 (i) Levels of diploid male production 314 Diploid males were heterozygous at a mean (range) of 5.5 (2-9) loci. The total number of 315 colonies exhibiting DMP remained the same regardless of the minimum number of 316 heterozygous loci (one to three) used to assign diploid males (Table S7) according to whether the criterion for assignment was a minimum of one, two, or three 322 heterozygous loci, respectively (Table S7). As the estimate of total number of colonies 323 exhibiting DMP was unaffected, the identity of the colonies classified as DMP colonies was 324 also unaffected, and the effect on the estimated total number of diploid males was very small 325 (the totals being 50, 50, or 49 diploid males), the criterion of a minimum of two heterozygous 326 loci to assign male diploidy was retained.

327
Frequency of DMP in field-collected colonies: One of the 20 field-collected colonies (5%), 328 all of which produced diploid workers, exhibited DMP ( Figure 2a;  assessed due to degradation of tissue inside the abdomen. All three queens in which mating 350 status could be determined had stored sperm inside the spermatheca, indicating that they were 351 mated ( Figure S4). Therefore, unmatedness was not the cause of exclusively haploid male 352 production within these colonies. Nonetheless, as these five colonies did not produce any 353 diploid offspring, they were excluded from estimates of DMP frequency, and subsequently 354 the calculation of allelic diversity at the sex-determining locus (see below).

355
In total, 37 of the 107 field-collected queens reared at least one adult offspring (Table S3), 356 with nine colonies producing only workers, six colonies producing only males, and 22 357 colonies producing both workers and males. All 6 + 22 = 28 colonies producing males were 358 sampled for diploid males (by genotyping). Of the six colonies producing only males, five 359 were found to produce haploid males only (and were excluded as above) and one (colony 360 102) to produce a single diploid male (Figure 2b, 2c; Table S3). Hence, 32 colonies produced 361 diploid offspring (nine colonies that produced only workers, 22 colonies that produced 362 workers and males, and the male-only producing colony that produced a single diploid male).  Table S3). Overall, therefore, of the 32 colonies producing diploid offspring, five colonies 365 (15.6%) exhibited DMP. Hence D was estimated to equal 5/32 = 0.156.

366
The five DMP colonies comprised four of the 12 (33.3%) first-brood male producing colonies 367 and one of the 16 (6.3%) late-male producing colonies (Figure 2b, 2c; Table S3). The four 368 first-brood male producing colonies that exhibited DMP produced one first-brood male each 369 and in each case it proved to be a diploid male (Table S3,  producing colonies; Figure 2b), a mean (range) of 2.6 (1-7) first-brood males were produced, 374 all of which were haploid (Table S3).    37 reared at least one adult offspring. Of these 37 colonies, 9 produced only workers, 6 501 produced only males, and 22 produced both workers and males. All 28 male-producing 502 colonies were sampled for diploid male production (by genotyping of males) (Table S3). Of 503 the 22 colonies that produced both workers and males, six produced 'first-brood males' (i.e.

504
males that eclosed within one week of first worker production), and 16 produced 'late males' 505 (i.e. males that eclosed a week or more after first worker eclosion). Hence, 12 first-brood 506 male producing colonies (the six colonies that produced only males, plus the six that 507 produced first-brood males and workers) and 16 late male producing colonies were defined.

508
The two classes of colony were exclusive in terms of timing of male production (males in 509 first brood vs. no males in first brood), but five of the first-brood male producing colonies 510 went on to produce late males as well (i.e. males were produced in both the first and