Allozyme Evidence for Genetic Autopolyploidy and High Genetic Diversity in Tetraploid Cranberry, Vaccinium oxycoccos (Ericaceae)

by Gregory Mahy, Leo P. Bruederle, Bridget Connors, Michael Van Hofwegen, Nicholi Vorsa
Allozyme Evidence for Genetic Autopolyploidy and High Genetic Diversity in Tetraploid Cranberry, Vaccinium oxycoccos (Ericaceae)
Gregory Mahy, Leo P. Bruederle, Bridget Connors, Michael Van Hofwegen, Nicholi Vorsa
American Journal of Botany
Start Page: 
End Page: 
Select license: 
Select License






'Unit6 d'Ecologie et de BiogCographie, UniversitC catholique de Louvain, place Croix du Sud-5, B-1348 Louvain-la-Neuve, Belgium;
?Department of Biology, Campus Box 171, University of Colorado at Denvel; PO. Box 173364, Colorado 80217-3364 USA; and
'Blueberry and Cranberry Research and Extension Centel; Rutgers University, Agricultural Experiment Station,
Chatsworth, New Jersey 08019 USA

Polyploidy has been important in the evolution of angiosperms and may significantly affect population genetic diversity and structure. Nineteen isoenzyme loci were studied in diploid and tetraploid populations of Vaccinilm~ ~x?~coccos (Ericaceae), and the results are compared with data previously reported for the related V. rnacrocarporz. Diploid V. o,~ycoccosand V. macrocarpon were readily discriminated based on their allozymic variation. No evidence for fixed heterozygosity was found in tetraploid V. oxycoccos. In contrast, all polymorphic loci exhibited both balanced and unbalanced heterozygotes, with some individuals exhibiting a pattern consistent with the presence of three alleles. These results support an autopolyploid origin for tetraploid V. o.uycoccos. However, tetraploid V. ~x?~coccos possessed a suite of alleles not found in diploid V. oxycoccos; half of these alleles were shared with V. macrocarpon. This suggests that autotetraploid V. oxycoccos may have undergone hybridization with V. nzacrocarpon or that the autotetraploid retained the genetic variation present in an ancestral diploid species. Following theoretical expectations, proportion of polymorphic loci, mean number of alleles, and observed heterozygosity were significantly higher for the autotetraploid than for the diploid. Mean inbreeding (F,,)was similar for diploid and tetraploid V. oxycoccos. The latter exhibited population differentiation (F,,) exceeding both diploid species.

Key words: allozyme variation; Ericaceae; F statistics; genetic diversity; polyploidy; Vacciniurn ri1acrocarport; Vacciniun~ o.xy- coccos.

Polyploidy has long been recognized as a major source of can be produced: balanced heterozygotes (AAaa) and two evolution and diversification in angiosperms, with as many as types of unbalanced heterozygotes (Aaaa, AAAa) (Soltis and 50% of all taxa estimated to have had a polyploid origin Rieseberg, 1986). Autopolyploidy can result in diallelic, trial- (Grant, 1981). Two major types of polyploidy are distin-lelic, and tetra-allelic loci, as well. Traditionally, autopoly- guished. Allopolyploidy is thought to result from hybridization ploidy has been viewed as maladaptative and was considered between variously cross-sterile taxa, with subsequent chro- rare in natural populations of plants (Stebbins, 1947, 1950). mosome doubling. As such, allopolyploids contain two differ- However, recent molecular data have demonstrated an auto- entiated genomes. In contrast, evidence suggests that auto-polyploid origin for an increasing number of polyploids and polyploidy arises from chromosome doubling of gametes at have stressed that autopolyploidy is of significant evolutionary the diploid level within a single species (Bretagnolle and importance (Soltis and Soltis, 1993). Thompson, 1995). AutoPol~Ploids have fully cross-fertile Pro- Both allo- and autopolyploids have potentially more genetic

genitors (Thompson and Lumaret, 1992; hnfort et al., 1998). variation than their diploid progenitors. The presence of two

In addition to mor~hological and c~tological observations, different genomes enables allopolyploids to produce all of the genetic data are of Prime importance in distinguishing among

enzymes possessed by each parent, as well as novel hetero- the types of polyploidy (Stebbins, 1950; Roose and Gottlieb,

merit enzymes (Thompson and Lumaret, 1992). Theoretically, 1976; Soltis and Rieseberg, 1986). AlloPol~Ploids are expect- natural populations of autopolyploids should exhibit a greater ed to exhibit disomic inheritance and fixed heterozygosity due level of heterozygosity than populations of their diploid pro- to the preferential pairing of divergent genomes. In contrast, genitor due to polyso~c inheritance, addition, polysom~c autotetraploids should be characterized by tetrasomic inheri- inheritance would enable an autopo~yp~oi~individual to havetance, or a higher level of ~ol~somic inheritance in the case more than two alleles at a single locus, As a result, an auto- high-1eve1 autO~O1~~lOids. in the Presence This polyploid plant could produce more heterodimeric enzymes

both and heteroz~gOtesin auto~ol~- than a heterozygous diploid individual, This increase in inter- plaids. when (A and n) are present nal genetic variation is thought to be a key element in the

at a locus in an autotetra~loid.three of heteroz~gotes success of polyploid cytotypes (Levin, 1983; Thompson and Lumaret$ 1992; and Felber$1993)'

I Manuscript received 22 April 1999; revision accepted 10 February 2000.

The authors thank the University of Colorado at Denver for administrative PO1~~lOid~ only influences the level genetic support of G. Mahy (Research Associate) while conducting this research at in natural plant populations, but is also expected to affect the CU-Denver (G. Mahy also benefited from a "Bourse de recherches post- way this diversity is partitioned within and among populations. doctorales" and financial support from the UniversitC catholique de Louvain);

autotetraploid populations, equilibrium frequencies under

M. Jorgensen, who provided valuable data on unreduced pollen production in

random mating are attained Over genanonymous reviewers for their comments. erations, not in a single generation as is true for randomly

V. macrocarpon; 0.Hardy for comments and statistical advice; and three

Author for reprint requests (e-mail: mating diploid populations (Li, 1976; Bever and Felber, 1993).



This can result in more deviations from Hardy-Weinberg ex- pectations within populations of an autotetraploid. For demo- graphically similar populations, effective population size is doubled in an autotetraploid as compared to a diploid. As a result, the effect of genetic drift at equilibrium is reduced in autotetraploids as compared to expectations for diploids (Ron- fort et al., 1998). Although there is a growing set of empirical evidence that autotetraploids generally harbor more genetic variation than their diploid relatives (Soltis and Soltis, 1993), investigations of the partitioning of this genetic diversity among populations have been precluded by a lack of adequate analytical procedure. Only recently have these procedures been developed (Ronfort et al., 1998).

Documenting the origin of polyploidy in natural populations of plants and assessing its genetic consequences is of prime importance if we want to fully understand the reasons for the evolutionary success of polyploids. This study uses allozyme markers to address these questions in Vaccinium section Oxycocczls. Although widely recognized as comprising diploid, tetraploid, and hexaploid members, the taxonomy of Vaccinium section Oxycoccus remains unsettled. The most recent treatment of the section recognized only two species: V. ma- crocarpon Ait. and V. oxycoccos L. (Vander Kloet, 1983). Vaccinium macrocarpon, which is exclusively diploid (2n = 24), is endemic to North America, with its natural distribution extending from Newfoundland, west throughout the Great Lakes Region to Minnesota, and south through the Appala- chian Mountains to North Carolina and Tennessee. In contrast.

V. oxycoccos is a polyploid complex comprising diploids (2n = 24), tetraploids (2n = 48), and hexaploids (2n = 72) (Jac- quemart, 1997). Whereas Vander Kloet (1983) recognized a single species, other authors considered diploids and tetra- ploids as morphologically distinct species, distinguishing the diploid V. microcarpum (Turcz.) Hooker fil, ex Rupr. & Schmalh. and tetraploid V. oxycoccos L. The hexaploid has either been called V, hagerupii (L. & L.) Ahokas or included in V. oxycoccos L. (Camp, 1944; Ravanko, 1990; Jacquemart, 1997). In this paper, we follow the classification of Vander Kloet (1983) and refer to V. oxycoccos sensu lato including diploids, tetraploids, and hexaploids. Vaccinium oxycoccos is a circumboreal species that is widespread across the north hemisphere. In North America it extends southward to Oregon in the Cascades and Virginia in the Appalachians. Diploid pop- ulations are limited to Alaska, extending south into Alberta and east to Hudson Bay. Tetraploid populations are more wide- spread, ranging from southern Alaska eastward to Great Slave Lake and Labrador, south to Northern California, Northern Idaho, Minnesota, the Great Lakes region, and New Jersey (Camp, 1944). According to Porsild and Cody (1980), tetra- ploids are very rare in the Northwest Territories where the diploids dominate. Although hexaploid V. oxycoccos has been documented in northern Europe, where it may be locally wide- spread (Ahokas, 1995), and in Asia; it's occurrence in North America, where it was originally documented by Newcomer (1941), has been postulated as sporadic (Camp, 1944). As such, hexaploid V, oxycoccos was neither considered, nor ev- idenced, in this study. Studying genetic variation in Vaccinium section Oxycoccus is of particular interest due to the impor- tance of V. macrocarpon as a small fruit crop species. An understanding of genetic variation in the related V. oxycoccos may enhance breeding opportunities with V. macrocarpon.

Tetraploidy in Vaccinium section Oxycoccus has been eval- uated on the basis of morphology and cytology (Camp, 1944;

TABLE1. Location of four tetraploid Vnccinium oxycoccos (04n), four diploid V. oxycoccos (02n), and ten V. nzacrocarpon (M) popula- tions sampled for the study of allozyrne variation. N = sample size.


Population Slte, state North latitude West longitude N

Tetraploid Vaccinium oxycoccos 04n-1 Hartwick Pine, MI 04n-2 Ryerse Lake, MI 04n-3 Dollar Lake, MI 04n-4 Cedarburg Bog, WI

Diploid V. oxycoccos 02n-1 Highway fout; AK 02n-2 Cantwell, AK 02n-3 Little Nelchina Rivet; AK 02n-4 Talkaetna, AK

V. macrocarpon

M-l Dow Lake, MI
M-2 Island Beach, NJ
M-3 Ryerse Lake, MI
M-4 Fish Lake, MI
M-5 Pilgrim Lake, MA
M-6 Cranberry Lake, MI
M-7 Cape Henelopen, DE
M-8 Vermillion, MI
M-9 Arbor Vitae, WI
M-10 Mahoning, PA

Vander Kloet, 1983; Ravanko, 1990). Camp (1944) considered the diploid V. macrocarpon to be primitive within the section and the circumboreal V. oxycoccos to be derived; initially, the latter was assumed to be entirely diploid. Camp (1944) pro- posed that allotetraploidy resulted from chromosome doubling in hybrids that were formed when the range of diploid V. ox- ycoccos was pushed southward by glaciation, rendering it sym- patric with V. macrocarpon. In contrast, an autopolyploid or- igin would be consistent with the subtle morphological differ- ences between diploid and tetraploid V. oxycoccos reported by Vander Kloet (1983), as autotetraploids are typically assumed to resemble closely the parental diploid species (Grant, 1981).

In this study, we assess isozyme variation in diploid and tetraploid populations of V. oxycoccos and compare these data with results reported for V, macrocarpon (Bruederle et al., 1996). This paper attempts to test the following hypotheses:

diploid V. oxycoccos and V. macrocarpon are genetically distinct; (2) tetraploid V. oxycoccos has had an autopolyploid origin (Vander Kloet, 1983); and (3) tetraploid populations of
oxycoccos exhibit a higher genetic diversity than their dip- loid counterparts (Levin, 1983; Thompson and Lumaret, 1992; Bever and Felber, 1993), and the partitioning of genetic di- versity among populations differs between cytotypes (Ronfort et al., 1998).

Species arid samplirrg-The species comprising section Oqcocc~rsare trailing vines with slender flexible, woody branches. Vaccinium oqcoccos has ovate leaves, the largest being <l cm, and pedicels bearing red scale- like bracts <1 mm wide; this contrasts with V. macrocarpon, which possesses narrowly elliptic leaves, the largest >1 cm, and pedicels bearing leaf-like bracts >1 mm wide (Vander Kloet, 1983).

Four diploid populations of V. oxycoccos were sampled in Alaska, while four tetraploid populations of V. o,~~~coccos

were sampled in Michigan and Wisconsin (Table 1). Because tetraploids were reported previously to be very rare in Alaska (Porsild and Cody, 1980); no search specifically for the tetra- ploid cytotype was made in this region. Chromosome counts were obtained from five putatively diploid individuals collected in Alaska to confirm the cytotype (Nick Vorsa, unpublished data). The mean geographic distance among populations of each cytotype was 172.9 km (SD = 52.3) and 235.8 km (SD = 141.3) for the diploids and tetraploids, respectively. Ramets were collected as stem sections sampled at 1-3 m intervals along parallel transects dissecting the sites. In Alaskan populations, care was taken to sample indi- viduals exhibiting the typical morphology of diploid V. onycoccos. According to Camp (1944) and Ravanko (1990), diploids exhibit smaller leaves (2-6 mm long, 1.2-2 mm wide) than the tetraploids (2-10 mm long, 1.5-5 mm wide). Additional characters, including the size of bracts borne on the pedicels and the pubescence of the pedicel, were used for dubious individuals (Camp, 1944; Ravanko, 1990). Genotypic data collected by Bruederle et al. (1996) for ten populations of V. rnacrocarpon were used for comparison with V. ox?Jcoccos(Table 1).

Electrophoresis-Extraction and electrophoretic methods followed Brued- erle et al. (1996). Thirteen enzymatic systems were studied with three different starch gel (10.5%) and electrode buffer systems. A lithium borate pH 7.618.0 system was stained for aspartate-amino-transferase (AAT), leucine-amino- peptidase (LAP), alcohol dehydrogenase (ADH), glucose-6-phosphate isom- erase (GPI), superoxide dismutase (SOD), and triose-phosphate-isomerase (TPI); a histidine-HC1 pH 7.0 system was stained for glyceraldehyde-3-phos- phate dehydrogenase (G3PD), menadione reductase (MNR), 6-phosphoglu- conate dehydrogenase (PGD), phosphoglucomutase (PGM), and shikimic acid dehydrogenase (SDH); and a morpholine-citrate pH 6.1 system was stained for isocitrate dehydrogenase (IDH) and malate dehydrogenase (MDH). For diploid populations, interpretation of variable enzyme patterns was based on the known subunit composition and number of isozymes commonly observed in diploid plants, as well as inheritance patterns reported previously for the related Vacciniunz section Cvanococcus (Vorsa, Manos, and Van Heemstra, 1988; Van Heemstra, Bruederle, and Vorsa, 1991). For each tetraploid plant, banding patterns were examined for relative band intensities that were inter- preted as corresponding to genotypes of different allelic dosage. Most variable tetraploid plants were electrophoretically examined more than once to ensure accurate assignment of genotype. Standards obtained from individuals known to be heterozygous and homozygous for each taxon, were used to help verify genotypes. The nomenclature of loci and alleles followed Bruederle et al. (1996). Alleles were reassigned following identification of new alleles in dip- loid and tetraploid V. oxycoccos.

Analysis-To provide insight into the origin of polyploidy in cranberry, patterns of enzyme variation were examined for the presence of balanced and unbalanced heterozygotes, as well as for evidence of fixed heterozygosity. The number of alleles in each species at each ploidy level was enumerated. The number of alleles shared by each taxon, as well as the number of alleles unique to tetraploids was also quantified. Principal components analysis (PCA) was performed using presence (1) or absence (0) for each of 36 alleles in 74 tetraploid V. oxycoccos individuals, 96 diploid V. onycoccos individuals, and 274 V. nzacrocarpon individuals using SYSTAT (1989). Loci monomor- phic across all populations and alleles present in all individuals were not included in the analysis. Data for the Pgi-I locus were omitted because, in diploid V. oxycoccos populations, genotypes at this locus were collected from a set of individuals different from those for the other loci. Individuals from the tetraploid population 04n- 1 were not included in this analysis, as we were not able to resolve Adh-I and Sdh-I in this population. From the data matrix, a correlation matrix was derived from which principal components were ex- tracted. Relationships among individuals were visualized by projecting indi- vidual scores onto the planes defined pairwise by the principal components.

The mean number of alleles per locus (A), percentage of polymorphic loci at the 0.05 criterion (P), and observed heterozygosity (H,) were calculated for each population from allozyme data using Biosys-1 (Swofford and Selan- der, 1981) for diploid populations, and by hand for tetraploid populations. In the tetraploid population 04n-4, estimates were based on 17 loci instead of 19, as Adh-I and Sdh were not resolved in this population. Population values for A, P, and H, did not exhibit skewness and kurtosis, indicating no deviation from a normal distribution. Pairwise t tests were conducted to determine whether diploid V. ~x?~coccos, tetraploid V. ~x?~coccos, and V. rnacrocarpon differed significantly for the aforementioned population genetic parameters.

Furthermore, inbreeding within populations and the level of differentiation among populations were evaluated using F statistics (Wright, 1951). The glob- al deficit of heterozygotes (FIT) was partitioned into its two components: the deficit of heterozygotes within populations (F,,) and the deficit of heterozy- gotes among populations (F,,). For the diploid populations, the F statistics and their 95 and 99% confidence intervals (CI) were estimated using F-STAT (Goudet, 1995) following the method of Weir and Cockerham (1984). Assum- ing that the tetraploid behaves genetically as an autopolyploid (see Results and Discussion), F statistics for these populations were estimated using GENE4X, following the modification for autotetraploidy of the method of Weir and Cockerham (1984) by Ronfort et al. (1998). F statistics were esti- mated with all variable loci (Table 2) in each set of populations. For the tetraploid, A&-I and Sdh-1 were not included in the data set, as these loci were not revealed in population 04n-4. Because FsTestimates in autotetra- ploids are expected to vary across loci as a consequence of different amounts of double reduction during meiosis, we also used the function of identity probabilities, p, defined by Ronfort et al. (1998) to assess population differ- entiation. The parameter p can be used to assess population structure over many loci without any prior knowledge concerning the proportion of double reduction. For tetraploids, p was estimated with GENE4X. For diploids, we approximated p by the formula (Ronfort et al., 1998):


a = [1(K -l)F,,]IK and K = 2


Enzyme electrophoresis resolved 19 putative loci represent- ing 11 enzymatic systems in the three taxa: Adh-I, Adh-2, G3pd-2, Gpi-I, Gpi-2, Idh-2, Mdh-I, Mdh-2, Mdh-3, Mdh-4, Mnr-3, Pgd-2 Pgm-1, Pgnz-2, Sdh-1, Sod-1, Tpi-1, Tpi-3, Tpi-

4. AAT and LAP were too faint to be scored in diploid V. oxycoccos. Despite good results previously obtained for tet- raploid V. oxycoccos and V. macrocarpon, these enzymes were omitted from the subsequent statistical analyses. Five of the 19 loci were monomorphic across all populations of the three taxa: Mdh-I, Mdh-4 Tpi-1, Tpi-3, and Tpi-4. Although Pgm-I was monomorphic within taxa, populations of V. macrocarpon and V. oxycoccos were fixed for alternate alleles (Table


Those individuals identified as diploid by morphological ex- amination expressed the expected diploid isozyme patterns. They were either homozygous or balanced heterozygotes. Electrophoretic banding patterns concurred with expectations for each of the enzymes examined. However, banding patterns for Adh-1 and Adh-2 both dimeric enzymes, were complicated by the formation of heterolocus dimers. In addition, the fastest allele at Adh-2 (i.e.,Adh-2a) comigrated with the slowest allele at Adh-1 (i.e., Adh-lb). This resulted in atypical staining, be- cause of the comigration of Adh-2 heterodimers with hetero- locus forms.

Tetraploids expressed more complex patterns. Dosage ef- fects were common, with unbalanced heterozygotes at all poly- morphic loci in these populations. At Pgi-2 a dimeric enzyme, two banded patterns were occasionally observed, with one band more intense than the other. This was interpreted to be a three-banded unbalanced heterozygote in which the faintest band could not be distinguished. At this locus, some tetraploid individuals displayed a five-banded pattern that was not found in the diploids. The second, third, and fourth bands were equi- distant, corresponding to Pgi-2b, Pgi-2c, and Pgi-2d, respec

Z%Z 8 8 2%%%?Z E 8 $8 8 8 $2; ca

I lo?? 9 9 9"?9F" 9 0-82 2 -9 zzg 9


OOC -icocoa0

CC a00 C COaOCOCC 8 gg $2 gggg oaooa

NW CON0 0 Q-PI-ocmm

N" "No 9 9%""""9? Nr g8 S9MS qa~rs


852: ss 8

a0 coo--cocoa0 aoaoo oaoa CO i 2% 050 c oaoaoaa 0 oaoc 0 ooc co 0

a c ~N~CONNW a d m c mi -a c

"M YN9 C =?"9"?9? 9 2 % q 228 Z$ q

a0 00-* OCOOaOa i

omt--o mi a c a C C CO 0 C

CiCC t-NOD 0 Nr-C 0 a a 0 mln 0 0

g 22 q -9 =9 9"C! 9 0

ii -co -i OQd COO 00 a g g c mm a a

omw coc 8 aa a a me 0 a

sgz 222 i9 09 9 9 0 9 $8 99

i-4 i i i i

000 CCdS 0 0 000 C0a0

CNM WNdln 0 Nee 8 8 8 8 dNeO

222 2 q 8 8 9s~o c: 9 9 1"-9

--OCO ---i OCO

000 oac 0

ado aao c 8 8 q88 88 $8 8

482 222 9 c 9

ii zc:ss --22 =


8 %$go0 8 %% 8 8 oa 8 8 8 g%g 2% 8

9 "t-99 9 gz g 9 9 9 9 9 SF9 22 g

-coo--4 iiiOCO


0 COO C t-m c

8 Z,"G8 8 cc~or-NC oa 0 O OC 0 m~ c

4 2225 9-282 9-9 =99 = 28 5

i --i

0 Ncc a 000 0 r-m a

o m~ 8 8 :3%8 8 0 m~

000 0 a

9 "09 9 9 """9 9-9 999 -9 N" 9

4 a0 -* oao+ i -4-00

0 00 0 0 COC 0 CO C C

o co c 88 8 0 clnln 0 0

9 99 = 99 = 9 q 5% 9 9

i i---i i i -i

Fig. 1. Principal components analysis based on the presence of alleles in individuals of diploid Vaccirzium oxycoccos (02n), tetraploid V. oxycoccos (04n), and V. macrocarpon (M). The first two components accounted for 31% of the total variation.

tively. Since Pgi-2c is intermediate in mobility between Pgi- 2b and Pgi-2d, it comigrates with the heterodimer formed by these alleles. Tetraploid plants possessing these three alleles therefore exhibited a five-banded pattern. Such an interpreta- tion has already been proposed for Pgi-2 in the autotetraploid Tolmiea menziesii (Soltis and Rieseberg, 1986). Thirteen of the 106 (12.1%) tetraploid plants examined for this locus pos- sessed three alleles. Significantly, there was no evidence of fixed heterozygosity in the tetraploid plants at any of the loci examined. A small percentage of individuals (four out of 100) exhibited an unbalanced heterozygote pattern at the Adh-2 lo- cus. This confirmed the field observation (L. P. Bruederle, per- sonal observation) that some Alaskan populations of V. oxycoccos contained a mixture of cytotypes, that is, both diploids and tetraploids. These individuals were removed from the data set.

A total of 46 alleles was found at the 19 loci (Table 2). Whereas diploid V. oxycoccos and V. macrocarpon possessed 28 and 31 alleles, respectively, tetraploid V. oxycoccos pos- sessed 40 alleles. Twenty alleles were common to both taxa. Tetraploid V. oxycoccos exhibited 15 alleles that were not found in diploid V. oxycoccos; eight of these alleles were shared with V. macrocarpon, while seven were unique to tet- raploid V, oxycoccos. In some cases, diploid V. oxycoccos and

V. macrocarpon shared the same predominant allele at poly- morphic loci: Adh-I, G3pd-2 (excepting population 02n-2), Mdh-2, Mdh-3, Mnr-3, Pgd-2, Pgm-2, Sdh-I, and Sod-I. At these loci, tetraploid populations of V. oxycoccos also exhib- ited the same predominant allele as their diploid counterpart, except for Pgd-2 in the 04n-2 population. At loci where dip- loid V. oxycoccos and V. macrocarpon differed for the pre- dominant alleles, the tetraploid populations shared (1) the pre- dominant allele with diploid V. oxycoccos (Pgm-I), (2) the predominant allele with V. macrocarpon (Pgi-I), or (3) the predominant allele with either diploid V. oxycoccos or V. macrocarpon, depending on the population (Gpi-2, Idh-2). In some cases, the predominant allele in the tetraploid popula- tions was not the most common in any of the diploid popu- lations, for example, populations 04n-2 and 04n-4 for Adh-2 and population 04n-2 for Pgd-2.

The principal components analysis readily discriminated the three taxa (Fig. 1). The first component, which accounted for

TABLE3. Population genetic diversity estimates within tetraploid Vaccirliurn oxycoccos (04n), diploid V. oxycoccos (02n), and V. nza- crocarpon (M). N = number of individuals sampled; P = proportion of polymorphic loci; A = mean number of alleles per locus; H, = proportion of observed heterozygosity. Means i 1 SD are given. Means with the same letter do not differ significantly O, <

0.05) based upon a t test.
Genetic d~versity
Population    P    4    Ho
Tetraploid V. oxycoccos            
Mean genetic diversity            
Diploid V. oxycoccos            
Mean genetic diversity            
V. nzacrocarpon            
M- 1            
Mean genetic diversity            
21% of the variance, clearly separated diploid V. oxycoccos from V. macrocarpon. Although tetraploid V. oxycoccos was also separated from V. macrocarpon, it partially overlapped diploid V. oxycoccos on the first component; the range of scores for tetraploid plants on this axis was intermediate be- tween the range of scores for the two diploid taxa. The second component, which accounted for 10% of the variance, clearly separated tetraploid V. oxycoccos from diploid V. oxycoccos and V, macrocarpon. The third component accounted for only 6% of the variance and was not considered further.

Genetic diversity estimates (P, A, H,) are displayed in Table 3 for each population and are averaged over V. macrocarpon, diploid V. oxycoccos, and tetraploid V. oxycoccos populations. The genetic diversity estimates were recalculated for V. macrocarpon, excluding the data for AAT and LAP. Averaged genetic diversity for tetraploid V. oxycoccos (P = 38.9, A = 1.66, H, = 0.213) was significantly higher (p < 0.01) than for either diploid V. oxycoccos (P = 15.8, A = 1.27, H, = 0.069) or V. macrocarpon (P = 11.1, A = 1.17, H, = 0.047). Higher mean heterozygosity in tetraploids, as compared to their diploid counterpart, may result from the presence of more polymorphic loci in tetraploids. In order to remove this effect, a single-locus comparison of observed heterozygosity was made at those loci for which both the diploid and tetraploid

V. oxycoccos exhibited variation. At each locus, average num- ber of alleles and average observed heterozygosity across pop- ulations were compared using a Mann-Whitney U test (Table 4). At four of the five loci examined, the tetraploid populations


TABLE4. Number of alleles (A) and observed heterozygosity (H,) averaged across populations at those loci for which diploid and tet- raploid V. oxycoccos were variable. * denotes significant differ- ences (p < 0.05) between cytotypes using the Mann-Whitney U test across populations.

A            H"
Locus    Diploid    Tetraplo~d        Diplo~d    Tetraploid
Aclh-2    2.0    2.5        0.408    0.598    *
Idh-2    1.25    2.00    *    0.088    0.671    *
Mdh-3    1.75    1.50        0.096    0.042    
Pgm-2    1.5    3.00    *    0.116    0.538    *
Sdh    1.75    2.67        0.130    0.339    
exhibited observed heterozygosity higher than their diploid counterpart; three of these differences were statistically sig- nificant (p < 0.05). It must be noted that, in general, tetraploid populations also exhibited significantly more alleles than dip- loid populations. At Idh-2 and Pgm-2, where tetraploid pop- ulations have a significantly higher observed heterozygosity, the difference in the number of alleles was also significant (p < 0.05) (Table 4).

Mean F,, values were low in the two diploid taxa and were not significantly different from zero, as indicated by their 95% CI (V. macrocarpon: F,, = -0.067, 95% CI = (-0,122, 0.083); diploid V. oxycoccos: F,, = 0.030, 95% CI = (-0.1 10, 0.275)). In tetraploid V. oq~coccos, mean F,, was also low (0.1 12). This value was included in the 95% CI for diploid V. oxycoccos, indicating that the mean level of inbreed- ing within populations was not significantly different between the two cytotypes. Mean F,, was 0.217 (99% CI = (0.142, 0.285)) for V. macrocarpon and 0.147 (99% CI = (0.0260.246)) for V. oxycoccos, indicating a low but significant level of differentiation among diploid populations. Mean F,, for tet- raploid populations (0.297) was twice the mean value for dip- loid V. oxycoccos. Nevertheless, in the tetraploid, F,, values were quite variable among loci, ranging from 0.023 (Mdh-2) to 0.835 (Pgd-2). At 0.835, the value of F,, at Pgd-2 was particularly high. When this locus was excluded from the anal- ysis, mean F,, for the tetraploid (0.201) was still higher than for diploid V. oxycoccos. Estimates of the parameter p also indicate a higher genetic differentiation among tetraploid pop- ulations (p = 0.558), when compared to diploid V. oxycoccos (p = 0.251) and V. macrocarpon (p = 0.372).


This study provides genetic evidence supporting autopoly- ploidy with high levels of genetic diversity in tetraploid pop- ulations of V. oxycoccos. However, considering the circum- boreal distribution of V. oxycoccos and the potential for mul- tiple origins of polyploidy (Thompson and Lumaret, 1992; Soltis and Soltis, 1993), care should be taken when extrapo- lating these results across the range of this widespread taxon.

The two diploid cranberry species, both potential progeni- tors of tetraploid V. oxycoccos, are readily distinguished based on isozyme variation. Principal components analysis based on presencelabsence of alleles clearly separated diploid V, oxycoccos from V. macrocarpon along the first principal com- ponent. Because of this differentiation, true allopolyploidy in tetraploid V. oxycoccos would have been expected to result in fixed heterozygosity at some of these loci, as allopolyploids represent a permanent hybrid possessing two divergent ge- nomes (Roose and Gottlieb, 1976; Soltis and Rieseberg, 1986). At the Pgm-I locus, where the two diploid taxa are fixed for different alleles, an allotetraploid would have been expected to display fixed heterozygosity. In tetraploid V. oxycoccos, plants exhibited homozygous, as well as heterozygous pat- terns, at all polymorphic loci.

In contrast, patterns of variation observed at polymorphic isoenzyme loci substantiate tetrasomic inheritance. Both bal- anced and unbalanced heterozygotes were observed. At the most variable locus, Pgi-2, some individuals exhibited a pat- tern consistent with the presence of three alleles. Tetrasomic inheritance is the most likely explanation for the observation of both balanced and unbalanced heterozygotes at polymorphic loci and the complex patterns observed among heterozygous individuals at the Pgi-2 locus. Because tetrasornic inheritance is essential for a diagnosis of autotetraploidy, we plan to assess it directly in the future using controlled crosses. The propor- tion (1 2.1 %) of tetraploid individuals possessing three alleles is low when compared to those observations reported for other tetraploid species. For example, Soltis and Soltis (1989) re- ported that 39% of autotetraploid Tolmiea plants that were examined possessed three or four alleles at one or more loci, with a proportion of up to 71% in some populations. In V. oxycoccos, three or four alleles were found in at least one tetraploid population for only five of the ten polymorphic loci: Adh-2 Pgi-2, Mdh-3, Pgm-2, and Sdh. In most cases (Mdh-3, Pgm-2 and Sdh), the most common allele occurred at a high frequency (>0.80), so rare alleles had a high probability of forming a heterozygote exclusively with the predominant al- lele. At Adh-2, we may have missed the presence of more than two alleles in some individuals because of the complex band- ing pattern due to the formation of heterolocus dimers. It is worth noting that a survey of tetraploid plants at Lap-1 also revealed a pattern consistent with the presence of three alleles at this locus, i.e., three-banded heterozygotes for a monomeric enzyme (L. I?Bruederle, unpublished data).

Morphological observations are also useful in distinguishing autopolyploidy from allopolyploidy. Morphologically, an au- totetraploid is typically assumed to resemble closely the pa- rental diploid species (Grant, 1981; Soltis and Rieseberg, 1983). The most recent treatment of section Oxycoccus (Van- der Kloet, 1983) revealed a clear separation of V.macrocarpon and diploid V. oxycoccos based on morphological characters. In contrast, morphological differences between diploid and tet- ra~loidV. oxvcoccos were more subtle. Vander Kloet (1983) his reported dontinuous morphological variation within'the oxycoccos polyploid complex, while Ravanko (1990) has shown that recognition of cytotypes morphologically is pos- sibly due only to the variation in size of certain vegetative traits, which is greater in polyploids. Nevertheless, tetraploid forms that are intermediate between diploid 'oxycoccos and

macrocarpon have been reported in North America (Camp, 1944), suggesting that V. macrocarpon may have played a role in the evolution of some tetraploid cranberry populations.

Genetic autopolyploids are expected to exhibit a high ge- netic similarity with their diploid progenitor. In fact, diploids and autotetraploids have been shown previously to possess es- sentially identical suites of alleles at all loci, as well as several predominant alleles in common (Ness, Soltis, and Soltis, 1989; Soltis and Soltis, 1989; Gauthier, Lumaret, and BCdCcarrats, 1998). In V. oxycoccos, the allozymic data did not fit expec- tations, that is, high genetic similarity between autotetraploids and their putative diploid progenitor. Not only did tetraploid V. oxycoccos possess 15 alleles not found in diploid V. oxy- coccos, of which eight were shared with V, macrocarpon, but in a significant number of cases, tetraploid V. oxycoccos shared predominant alleles with V. macrocarpon and not with diploid

V. oxycoccos. Finally, the principal components analysis clear- ly separated diploid and tetraploid V. oxycoccos along the sec- ond component. Two sets of hypotheses, which are not mu- tually exclusive, may help to explain a genetic similarity that is lower than expected between diploid and tetraploid V. ox- ycoccos.

First, despite the fact that patterns of allozymic variation support tetrasomic inheritance, we cannot rule out definitively the possibility that tetraploids arose originally by hybridization between diploid V. oxycoccos and V. macrocarpon, in the re- gion investigated. Crosses between V. macrocarpon and Alas- kan diploid V. oxycoccos have resulted in excellent fruit set and seed production comparable to intraspecific crosses, in- dicating that the species are cross-compatible (N. Vorsa, un- published data); however, chromosome counts to determine ploidy have not yet been performed on the progeny. While diploid V. oxycoccos and V. macrocarpon currently display parapatric distributions, they may have been in contact in the past, rendering interspecific hybridization possible (Camp, 1944). Diploid species may be differentiated based on mor- phological or allozyme variation and still exhibit low genomic differentiation (Qu, Hancock, and Whallon, 1998). Then, if the genome of the two diploids are combined in a tetraploid, chro- mosome homologies may be sufficient for pairing to be ran- dom among the two genomes leading to tetrasomic segregation patterns. In a recent study, Qu, Hancock, and Whallon (1998) suggested that little genomic divergence has developed be- tween Vaccinium species. Additionally, tetraploid V. oxycoc- cos was sampled primarily within the range of V. macrocarpon and, in some cases, populations of the two taxa were syntopic (Table 1). In this part of their range, autotetraploid V. oxycoccos may have undergone hybridization with V. macrocarpon, which could explain the observation of "macrocarpoid" tetraploid forms reported by Camp (1944). Because of the low genomic divergence between some diploid Vaccinium, polyploids may freely exchange genes with sympatric diploid spe- cies via unreduced gametes (Qu, Hancock, and Whallon, 1998). However, the potential mechanism of hybridization be- tween tetraploid V. oxycoccos and V. macrocarpon is unclear. Indeed, preliminary observations suggest an extremely low frequency of unreduced gamete formation in natural popula- tions of V. macrocarpon; unreduced pollen production ranged from 0 to 0.041% over the ten populations for which allozyme data were also collected (M. Jorgenson and L. P. Bruederle, unpublished data). Whether tetraploid V. oxycoccos resulted from hybridization between different species or arose within a single species, this does not preclude tetraploids behaving ge- netically as autopolyploids.

Alternatively, a reduction in the effect of genetic drift due to the inertia of the autotetraploid genetic structure may ac- count for the observation of the difference in genetic compo- sition observed between di~loid and tetradoid V. oxvcoccos. and for the presence of alleles unique to tetraploids. Regardless of whether V. oxycoccos or V. macrocarpon is more recently derived within section Oxycoccus, there is little doubt that both species have undergone a considerable shift in their distribu- tion during the course of glaciation. In North America, diploid

V. oxycoccos currently occupies regions that were, for the most part, previously glaciated. Shifts in the distribution of species during glaciation and recolonization of their current range have most probably resulted in a series of bottlenecks and founder events, which, in conjunction with genetic drift, can significantly decrease allelic diversity maintained by diploid populations (Nei, Maruyama, and Chakraborty, 1985). In con- trast, because tetraploids are less prone to genetic drift (Moody, Mueller, and Soltis, 1993; Ronfort et al., 1998), au- totetraploid V. oxycoccos may have retained most of the var- iation that was initially present in the ancestral diploid species of Vaccinium section Oxycoccus. Sampling of additional pop- ulations throughout the entire range of V. oxycoccos is needed to assess the importance of genetic drift and hybridization in tetraploids. It will be particularly interesting to compare ge- netic structure in diploid and tetraploid populations from the European continent, where V. macrocarpon does not naturally occur.

As has been reported for most autotetraploid taxa studied previously (Soltis and Soltis, 1993), tetraploid V. oxycoccos displayed statistically higher values for A, P, and H,in com- parisons with the diploid. Specifically, autotetraploids are ex- pected to display higher levels of heterozygosity than their diploid progenitors due to the effect of tetrasomic inheritance (Stebbins, 1980; Moody, Mueller, and Soltis, 1993). In V. ox- ycoccos, the increased mean heterozygosity observed in the tetraploid may also be the result of a greater proportion of polymorphic loci, although we have shown that tetraploids generally exhibited a greater proportion of heterozygotes at individual loci. Nevertheless, in addition to the effect of tet- rasomic inheritance, the higher heterozygosity observed in the tetraploid plants may be due, in part, to the fact that they possess numerous alleles not found in diploids and generally maintained more alleles at individual loci.

Both in Europe and North America, tetraploid V. oxycoccos occupies a broader range than its diploid counterparts. While it has been reported that tetraploid V. oq~coccos is very rare in Alaska (Porsild and Cody, 1980), the occurrence of indi- viduals with typical tetraploid isozyme patterns among our Alaskan populations indicated that the range of this cytotype is more widespread than previously thought. Wider ecological niches and successful colonization of a greater range of hab- itats in polyploids, as compared to diploids, have been inter- preted as a consequence of greater enzyme diversity (Levin, 1983). If there is a relationship between allozymes and the conditions under which they operate optimally, then the eco- logical success of polyploids should be related directly to their higher heterozygosity and enzyme multiplicity, resulting in an increase in distinct functional enzymatic forms, a greater ex- pression of overdominance, and potentially higher phenotypic plasticity (Stebbins, 1980, 1985; Levin, 1983; Bever and Fel- ber, 1993; Soltis and Soltis, 1993).

Few investigations have considered the amount of genetic variation among naturally occurring autopolyploid populations in comparison with the related diploid species (Ronfort et al., 1998), and data are needed to test the predictions of theoretical models. The genetic structure of tetraploid V. oxycoccos does not support the expectations that autotetraploid populations should differ from related diploid populations in two ways: (1) more deviations from Hardy-Weinberg equilibrium and (2) lower levels of population differentiation. First, F statistics in- dicated a low level of inbreeding, both in diploid and autotet- raploid populations of V. oxycoccos. Second, the mean level of population differentiation was higher or similar in tetraploid than in diploid V. oxycoccos.


AHOKAS, H. 1995 Is the polyploid cranberry (Vaccinium sp.) in Finland tetraploid or hexaploid? Nordic Jo~irnal of Botany 16: 185-189.

BEVER.J. D., AND E FELBER. 1993 The theoretical population genetics of autopolyploidy. Oxford S~irvey of Evolutionary Biology 8: 185-217. BRETAGNOLLE, 1995 Tansley Review number 78.

E, AND J. D. THOMPSON. Gametes with the somatic chromosome number: mechanisms of their formation and role in the evolution of autopolyploid plants. New Phy- tologist 129: 1-22.

BRUEDERI-E,L. I?, M. S. HUGAN, J. M. DIGNAN, AND N. VORSA. 1996 Ge- netic variation in natural populations of the large cranberry, Vaccinilim macrocarpon Ait. (Eiicaceae). Bulletin of the Torrey Botanical Cl~ib 123: 41-47.

CAMP, W. H. 1944 A preliminary consideration of the biosystematy of 0.xycocc~is. Bulletin of the Torrey Botanical Club 71: 426-437. GAUTHIER,P., R. LUMARET, AND A. BEDECARRATS.

1998 Genetic valiation and gene flow in alpine diploid and tetraploid populations of Lotus (L. alpinzis (D.C.) SchleicherlL. corniculatus L.). I. Insights from morpho- logical and allozyme markers. Heredity 80: 683-693.

GOUDET,J. 1995 Fstat v-1.2: a computer program to calculate F-statistics. Journal of Heredity 86: 485-486. GRANT, V. 1981 Plant speciation. Columbia University Press, New York, New York, USA. JACQUEMART,

A,-L. 1997 Biological flora of the British Isles: Vaccinizinl o,xycoccos L. (0x)~coccus palustris Pers.) and Vnccinilrm trzicrocarplrm (Turcz. Ex Rupr.) Schmalh. (0,xycoccus microcarpus Turcz. Ex Rupr.). Jourrzal of Ecology 85: 381-396.

KREBS, S. L., AND J. E HANCOCK.1981 Tetrasomic inheritance of isoenzyme markers in the highbush blueberry, Vaccinium cor~~mbosum

L. Heredity

63: 11-18. LEVIN, D. A. 1983 Polyploidy and novelty in flowering plants. American Naturalist 122: 1-25. LI, C. C. 1976 First course in population genetics. Boxwood Press, Pacific Grove, California, USA. MOODY, M. E., L. D. MUELLER, AND D. E. SOLTIS. 1993 Genetic variation

and random drift in autotetraploid populations. Genetics 134: 649-657. NEI: M., T. MARUYAMA, 1985 The bottleneck effect

AND R. CHAKRABORTY. and genetic variability in populations. Evolution 29: 1-10. NESS,B. D., D. E. SOLTIS,AND F! S. SOLTIS. 1989 Autopolyploidy in Heuchera rnicrantha (Saxifragaceae).American Journal of Botany 76: 614-


E. H. 1941 Chromosome numbers of some species and varieties of Vacciniwn and related genera. Proceedings of the American Society of Horticultural Science 38: 468-470.

PORSILD,A. E., AND W. J. CODY. 1980 Vascular plants of continental North-

west Temtories, Canada. National Museum of Natural Sciences, Ottawa,

Ontario, Canada.

Qu, L., J. E HANCOCK,AND J. H. WHALLON. 1998 Evolution in an auto- polyploid group displaying predominantly bivalent pairing at meiosis: genomic similarity of diploid Vacciniutn dnrrowi and autotetraploid V. coryn~bosum(Ericaceae). Ainericnrl Jo~irnal of Botany 85: 698-703.

RAVANKO,0. 1990 The taxonomic value of molphological and cytological characteristics in Oxjcoccus (subgenus of Vnccinium, Ericaceae) species in Finland. Annales Botanicea Fennici 24: 235-239.


AND E ROUSSET. 1998 Analysis of population structure in autotetraploid species. Genetics 150: 921-930. ROOSE, M. L., AND L. D. GOTTLIEB. 1976 Genetic consequences of poly- ploidy in Tragopogon. Evolution 30: 818-830.

SOLTIS, D. E., AND L. H. RIESEBERG. 1986 Autopolyploidy in Tolmiea ??/en- ziesii (Saxifragaceae): genetic insights from enzyme electrophoresis. American Jourtznl of Botany 73: 310-318.

, AND F! S. SOLTIS. 1989 Genetic consequences of autopolyploidy in Tollniea (Saxifragaceae). Evol~ition43: 586-594. --, AND --, 1993 Molecular data and the dynamic nature of poly- ploidy. Critical Reviews in Plarlt Sciences 12: 243-273. STEBBINS,G. L. 1947 Types of polyploids: their classification and signifi- cance. Advances in Genetics 1: 403-429. . 1950 Variation and evolution in plants. Columbia University Press, New York, New York, USA. --. 1980 Polyploidy in plants: unresolved problems and prospects. In

W. H. Lewis [ed.], Polyploidy: biological relevance, 495-520. Plenum, New York, New York, USA. . 1985 Polyploidy, hybridization and the invasion of new habitats. Annals of the Missouri Botanical Garden 72: 824-832.

SWOFFORD,D. L.. AND R. B. SELANDER. 1981 Biosys-1: a foltran program for the comprehensive analysis of data in population genetics and sys- tematics. Journal of Heredity 72: 281-283.

SYSTAT. 1989 SYSTAT Inc., Evanston, Illinois, USA.

THOMPSON,J. D., AND R. LUMARET. 1992 The evolutionary dynamics of polyploid plants: origins, establishment and persistence. Trends in Ecol- ogy and Evolution 7: 302-307.

VAN HEE~ISTRA, AND N. VORSA. 1991 Inheritance

M. I., L. P. BRUEDERLE, and linkage of thirteen polymolphic isozyme loci in diploid blueberry. Journal of the American Society of Horticzilt~iral Science 116: 89-94.

VANDERKLOET, S. P. 1983 The taxonomy of Vaccinium section Oxycoccus. Rhodora 85: 1-43. VORSA, N., F! S. MANOS, AND M. I. VAN HEE~ISTRA.

1988 Isozyme variation and inheritance in blueberry. Genome 30: 776-781. WEIR, B. S., AND C. C. COCKERHA~I.

1984 Estimating F-statistics for the analysis of population structure. Ellolution 38: 1358-1370. WRIGHT, S. 1951 Evolution in Mendelian populations. Genetics 16: 97-159.

  • Recommend Us