Mitochondrial DNA characterization of Africanized honey bee (Apis ...
Journal of Apicultural Research and Bee World 49(2): 177-185 (2010)
© IBRA 2010
DOI 10.3896/IBRA.1.49.2.06
ORIGINAL RESEARCH ARTICLE
Mitochondrial DNA characterization of Africanized honey bee (Apis mellifera L.) populations from the USA Allen L Szalanski1* and Roxane M Magnus1 1
Social Insects Genetics Lab, Department of Entomology, University of Arkansas, Fayetteville, AR 72701, USA.
Received 15 April 2009, accepted subject to revision 9 December 2009, accepted for publication 23 February 2010. *Corresponding author: Email: [email protected]
Summary We carried out a study which involved DNA sequencing of a portion of the mitochondrial DNA COI-COII region of Africanized honey bees (AHB) from the USA. A total of 12 mitotypes were observed, of which seven have not been previously described. Of the 172 samples, two mitotypes, A1 and A1d, accounted for 77% of the observed mitotypes, while mitotypes A1a, A26c, A26d, A29a, and A30 were only observed once. A possible reason why these new mitotypes have not been described before is because previous studies on AHB in the new world have relied primarily on PCR-restriction fragment length polymorphism (RFLP), which is less sensitive than DNA sequence data. Multiple mitotypes of ‘A’ lineage honey bees have previously been observed in South America and Mexico using PCR-RFLP and DNA sequence analysis. Our findings are consistent with previous studies of AHB genetic variation from central Mexico, Columbia, and northern Brazil, in that the A1 mitotype was more common than the A4. Maximum parsimony analysis revealed that all of the ‘A’ lineage Apis mellifera mitotypes formed a distinct clade relative to representatives of the ‘M’, ‘C’, and ‘O’ lineages. Statistical analysis of the mitotype frequencies in the USA revealed an excess of low frequency mitotypes, indicating that the population size is expanding. The amount of genetic variation observed in Africanized honey bees in the USA therefore supports the idea that there have been multiple introductions of AHB into the country.
Caracterización mitocondrial de poblaciones de abeja (Apis
mellifera L.) africanizada de los EEUU Resumen Se realizó un estudio mediante la secuenciación de una porción de la región COI-COII del ADN mitocondrial en las abejas de la miel africanizadas de los EEUU. Se observaron un total de 12 mitotipos, de los cuales siete no se habían descrito anteriormente. De las 172 muestras, dos mitotipos, A1 y A1d, representaron el 77% de los mitotipos detectados, mientras que los mitotipos A1a, A26c, A26d, A29a y A30 sólo fueron observados una vez. Una posible razón de por qué estos nuevos mitotipos no se hayan descrito antes es porque los estudios previos sobre la abeja africanizada en el nuevo mundo se habían basado principalmente en PCR-Polimorfismo de longitud de los fragmentos de restricción (RFLP), que es menos sensible que los datos de la secuencia de ADN. Varios mitotipos en las abejas de miel del linaje “A” habían sido previamente observados en América del Sur y México, utilizando PCR-RFLP y el análisis de secuencias de ADN. Nuestros resultados son consistentes con los estudios previos sobre la variación genética de la abeja africanizada en México central, Colombia y el norte de Brasil, en los que el mitotipo A1 es más común que el A4. El análisis de máxima parsimonia reveló que todos los mitotipos del linaje “A” de
Apis mellifera formaron un clado distinto en relación con los representantes de los linajes “O”, “M” y “C”. El análisis estadístico de las frecuencias de los mitotipos en los EEUU reveló un exceso de mitotipos de baja frecuencia, lo que indica que el tamaño de la población está en expansión. La cantidad de la variación genética observada en las abejas de miel africanizadas de los EEUU por tanto, apoya la idea de que ha habido introducciones múltiples de abejas africanizadas en el país.
Keywords: Apis mellifera, COI-COII intergenic region, Africanized honey bee, mtDNA, USA
Introduction
distributed across the world due to multiple migrations and introductions
Apis mellifera L. is native to Europe, Africa, and Asia (including Saudi
that are present in different regions of the world. These subspecies
Arabia, Iran and the Ural mountains of Russia). It is currently widely
have been classified into five main lineages: ‘C’ (the Carnica
(Ruttner, 1988). Apis mellifera includes about two dozen subspecies
178
Szalanski, Magnus
group that includes A. m. carnica and A. m. ligustica); ‘M’ (the north
al., 2003; 2004; 2007). In contrast to cyt b, the mtDNA cytochome
and western European honey bees that include A. m. mellifera,
oxidase I (COI) and COII genes for A. mellifera exhibit a high degree
A.m. iberica, and A. m. intermissa); ‘A’ (the African group that
of genetic variation within and amongst lineages, and is the preferred
includes A. m. scutellata, A. m. capensis, A. m. lamarckii, A. m.
molecular marker for detecting intraspecific genetic variation. This
litorea, A. m. adansonii, and A. m. unicolor); ‘Y’ (Ethiopia, Franck et
marker has been used for genetic analysis of honey bee populations
al., 2001); and the ‘O’ group (the Oriental or Middle Eastern group
from: Turkey (Solorzano et al., 2009); Mexico (Kraus et al., 2007);
which includes A. m. anatolica, A. m. caucasica, A. m. syriaca,
South America (Collet et al., 2006; Ferreira et al., 2009; Prada et al.,
A. m. pomonella, and A. m. cypria) (Ruttner, 1992). At the molecular
2009); Africa (Franck et al., 2001); and Australia (Chapman et al.,
level, these lineages are genetically divergent based on nuclear and
2008). Previous studies of COI-COII genetic variation of AHB in
mitochondrial DNA markers (Arias and Sheppard, 1996; Franck et al.,
Mexico, Columbia, Brazil and Uruguay have revealed several different
2001).
mitotypes of the ‘A’ lineage in each country, with two mitotypes
The Africanized honey bee (AHB) was first detected in Texas in
occurring in central Mexico (Kraus et al., 2007), six in Columbia
1990 (Sugden and Williams, 1990), and by 2009 had spread to a total (Prada et al., 2009), and five in Brazil and Uruguay (Collet et al., of ten states (Anonymous, 2009). The hybrid in the USA is virtually
2006). To date there is no information on COI-COII DNA sequence
indistinguishable in the field from the European honey bee (EHB), and variation of AHB in the USA, so the objective of our study was to requires a morphometric analysis for identification (Rinderer et al.,
determine the genetic diversity of AHB from the USA based on COI-
1993). Mitochondrial DNA (mtDNA) is an ideal genetic marker for
COII DNA sequence data.
identifying AHB, since a single worker can represent the entire colony (Sheppard and Smith, 2000). Introgression of AHB genes using a mtDNA marker is, however, not detectable if an EHB queen has mated with AHB drones.
Materials and methods Sampling
Previous studies of the molecular genetics of AHB in the USA were Specimens of adult worker honey bees were collected from Utah, New focused on molecular diagnostics, and primarily used a region of the
Mexico, Oklahoma, Texas, California, Florida and Arkansas and stored
mitochondrial DNA (mtDNA) genome, cytochrome b (cyt b) which has
in 70-100% ethanol until processed for DNA extraction (Table 1, Fig. 1).
a relatively low level of intraspecific variation in A. mellifera (Crozier et Samples were collected from managed colonies, feral colonies and
al., 1991), making it ideal for molecular diagnostic purposes (Pinto et
swarms. The two Arkansas samples were from two counties in the
Fig. 1. Frequency and locations of ‘A’ lineage Apis mellifera mitotypes from seven states.
Genetics of Africanized honey bees in the USA
179
Table 1. Distribution and frequency of the studied Apis mellifera ‘A’ lineage mitotypes from the USA. State
County
Lat/Long
Mitotype (n)
Arkansas (2)
Lafayette
33.26, -93.59
A1d(1)
Miller
33.32, -93.87
A1(1)
California (3)
San Diego
33.02, -116.77
A26a(1), A29a(2)
New Mexico (47)
Chaves
33.36, -104.47
A1(1), A1d(3), A26(2)
Curry
34.57, -103.35
A1(1), A1d(4), A26b(1)
De Baca
34.47, -104.24
A1(1)
Doña Ana
32.31, -106.77
A1(4), A1d(4), A26a(1)
Eddy
32.47, -104.3
A26a(1)
Grant
32.73, -108.38
A1d(1)
Guadalupe
34.86, -104.78
A1d(2)
Lincoln
33.74, -105.46
A1d(4), A26a(1)
Otero
32.62, -105.73
A1(1), A1d(4)
Roosevelt
34.02, -103.48
A1(1), A1d(4), A26a(1), A26b(3)
Santa Fe
35.51, -105.98
A1d(1)
Socorro
34.02, -106.93
A1(1)
Beckham
35.26, -99.69
A1d(1)
Blaine
35.88, -98.43
A1d(2)
Bryan
33.97, -96.25
A1(1)
Caddo
35.18, -98.38
A1(3), A4(1), A26a(1)
Carter
34.25, -97.29
A1d(1)
Cleveland
35.2, -97.33
A1(1), A1d(1)
Coal
34.6, -96.3
A1d(2)
Comanche
34.66, -98.46
A1(2), A1d(1)
Cotton
34.28, -98.37
A1(1), A1d(1)
Custer
35.64, -99.01
A1d(1), A26a(1)
Dewey
35.99, -99.00
A1d(3)
Grady
35.02, -97.89
A1(1)
Greer
34.93, -99.56
A1(1)
Harmon
34.74, -99.84
A1(1), A1d(1)
Hughes
35.04, -96.26
A1d(1)
Jefferson
34.10, -97.84
A1d(1)
Kiowa
34.92, -98.98
A1(1), A1d(7)
Love
33.95, -97.25
A1d(1), A26a(1)
Marshall
34.03, -96.77
A4(1), A26a(1)
McClain
35.00, -97.44
A1d(1), A26a(1)
McCurtain
34.11, -94.77
A1(1), A1d(1)
Murray
34.49, -97.07
A1(1)
Oklahoma
35.48, -97.53
A1(1), A26a(1)
Payne
35.48, -97.53
A1d(1)
Pottawatomie
35.20, -96.94
A1d(1)
Oklahoma (55)
180
Szalanski, Magnus
Table 1. Cont’d Distribution and frequency of the studied Apis mellifera ‘A’ lineage mitotypes from the USA. State
County
Lat/Long
Mitotype (n)
Stephens
34.48, -97.86
A1(1), A1d(1)
Tillman
34.38, -98.92
A1(1)
Washita
35.29, -98.99
A1d(1)
Washington
37.28, -113.52
A1(2), A1d(3), A1e(2), A26(1), A26a(1)
Kane
37.29, -111.89
A1d(2)
Iron
37.86, -113.28
A1(2), A1e(5), A26a(2)
Florida (1)
Broward
26.12, -80.24
A1d(1)
Texas (44)
Armstrong
34.97, -101.35
A1d(2)
Bastrop
30.10, -97.31
A1d(1)
Bexar
29.45, -98.52
A1(1)
Bosque
31.9, -97.63
A1d(1)
Calhoun
28.44, -96.61
A1(1)
Cameron
26.15, -97.45
A1(2)
Clay
33.79, -98.21
A1d(1)
Dawson
32.74, -101.95
A1d(1)
Dimmit
28.42, -99.75
A1d(1)
Fisher
32.74, -100.4
A1d(1)
Gaines
32.74, -102.64
A1d(1)
Gray
35.41, -100.81
A1d(2)
Gregg
32.48, -94.81
A1(1)
Guadalupe
29.58, -97.95
A29a(1)
Hamilton
31.70, -98.11
A1d(1)
Hidalgo
26.40, -98.18
A1(1)
Johnson
32.38, -97.36
A1d(1)
Jones
32.74, -99.88
A26a(1)
Kerr
30.06, -99.35
A1d(3)
Lee
30.31, -96.96
A1(1), A1a(1)
Lubbock
33.61, -101.82
A26(1), A30(1)
Matagorda
28.78, -96.00
A26c(1)
Pecos
30.78, -102.72
A1d(1)
Randall
34.97, -101.90
A1d(4)
Refugio
28.32, -97.17
A1d(1)
Swisher
34.53, -101.73
A1d(1)
Taylor
32.31, -99.88
A1d(1), A26a(1)
Travis
30.33, -97.78
A1d(1)
Uvalde
29.35, -99.76
A1d(1)
Victoria
28.8, -96.97
A1(1), A1d(1)
Williamson
30.32, -97.62
A1d(1)
Young
33.18, -98.7
A1(1)
Utah (20)
Genetics of Africanized honey bees in the USA
181
extreme south west corner of the state, and were collected by the
Mitochondrial analysis
Arkansas Plant Board in 2005 as part of their AHB monitoring
Genomic DNA from individual honey bee thoraces was extracted using
programme. The Oklahoma samples were collected in 2004-8
the Qiagen DNeasy extraction kit (QIAGEN; Valencia, CA, USA)
primarily from feral colonies and swarms as part of the AHB
according to the manufacturer’s protocol and per Solorzano et al.,
diagnostics programme conducted by Oklahoma State University.
(2009). A 637 to 1006 bp region of the COI-COII intergenic region
Identification of samples collected from New Mexico, Oklahoma and
was PCR amplified in a Techne T-412 thermal cycler (Techne Inc;
Arkansas from 2004-6 and from the Utah, Florida and California
Burlington, NJ, USA) using primers E2 and H2 (Garney et al., 1993).
samples was done using multiplex PCR following Szalanski and
PCR was conducted using a profile consisting of an initial denaturation
McKern (2007). Additional samples collected during 2007 and 2008
of 94oC for 2 min followed by 35 cycles of 94oC for 45s, 46oC for 45s,
from Oklahoma and New Mexico were determined as AHB using PCR-
and 72oC for 45s, and then a final extension of 72oC for 5 min.
RFLP (Pinto et al., 2003) by R A Grantham (Oklahoma State
Amplicons were separated using 1% agarose gel electrophoresis and
University, USA). Samples collected from Texas from 1991 to 2008
photo documented using a BioDoc-It™ Imaging System (UVP, Inc.;
were identified as Africanized (n = 26), Africanized with evidence of
Upland, CA, USA) per Solorzano et al., (2009). PCR products were
introgression of European genes (n = 11) or European with evidence
purified using Microcon-PCR Filter Units (Millipore; Bedford, MA, USA),
of introgression of Africanized genes (n = 6) using FABIS (Rinderer et
and sent to the University of Arkansas Medical Sciences DNA
al., 1993) by L Bradley (Texas A&M University, USA). Utah samples
Sequencing Core Facility (Little Rock, AR, USA) for direct sequencing
were collected in 2008 and 2009 by the Utah Department of
in both directions. DNA sequences new to this study were deposited
Agriculture and Food. Voucher specimens are deposited at the
in NCBI GenBank (http://www.ncbi.nlm.nih.gov/) as accession
Arthropod Museum, Department of Entomology, University of
numbers FJ743632 to FJ743641, FJ890929, FJ890930, GU326335 and
Arkansas, Fayetteville, AR, USA.
GU326336.
Fig. 2. Phylogenetic relationship among mitotypes representing four lineages of Apis mellifera based on an unrooted maximum parsimony heuristic analysis. Maximum parsimony bootstrap values (>50%) are provided, and GenBank accession numbers are provided for each mitotype.
182
Szalanski, Magnus
Genetic diversity analysis
Results
DNA sequences were aligned with CLUSTAL W (Thompson et al., 1994) using Bioedit v5.0.7 (Hall, 1999). Designation of mitotypes was done
Table 1 shows the DNA sequencing analysis of honey bee samples
by comparing sequences with those available on GenBank. Number of
collected from New Mexico, Utah, Texas, Florida, Oklahoma,
mitotypes and their frequencies were determined both visually and
California, and Arkansas. A total of 12 mitotypes were observed,
with the program DNAsp version 4.10.9 (Rozas et al., 2003). DNAsp
which ranged from 637 bp to 831 bp in size (Table 2). Four mitotypes
was also used to estimate the following variables: haplotypic diversity
have previously been described. Mitotype A4 was identical to GenBank
(Hd) (Nei, 1987), and Nei’s Nm value. Nucleotide diversity was
EF033650 from Brazil, mitotype A1 to EF033649 from Brazil, mitotype
interpreted as the average proportion of nucleotide differences
A30 to EF033654 from Brazil, and mitotype A26 to FJ477990 from
between all possible pairs of sequences in the sample (Hartl and
Namibia. Based on a BLAST search of GenBank DNA sequences, five
Clark, 1997), mean number of pairwise nucleotide differences (K)
of the 12 mitotypes observed have not previously been described.
equation A3 (Tajima, 1983), number of polymorphic sites (S), and the Mitotypes A1d and A1e were most similar to EF033649 mitotype A1 parameter Өg. The parameter θ is the proportion of nucleotide sites
(0.5% divergence), mitotype A29a was closest to EF033653 mitotype
that are expected to be polymorphic in any suitable sample from this
A29 (0.3% divergence) and mitotypes A26a to A26d were most
region of the genome (Hartl and Clark, 1997). To test for neutral
similar to FJ477990 mitotype A26 (1.6 to 1.8% divergence). Overall,
mutation, Tajima’s D (Tajima, 1989), and D* and F* (Fu and Li, 1993) mitotypes A1 and A1d were the most common, accounting for 77% of were calculated. To examine demographic stability, Fu’s Fs statistic
samples, while the rest of the 10 observed mitotypes accounted for
(Fu, 1997) (based on mitotype distribution) was used.
the remaining 23% of the samples.
For the phylogenetic analysis, DNA sequences were aligned using
In Arkansas, two AHB samples were found, one A1 and one A1d
CLUSTAL W (Thompson et al., 1994) using DNA sequences from this
mitotypes (Table 1, Fig. 1). In Oklahoma, a total of 55 samples from
study and additional ones from GenBank (Fig. 2). Maximum
28 counties were sequenced, with mitotype A1d being the most
parsimony (MP) analysis on the alignments was conducted using
common, followed by A1, A26a and A4. Mitotype A4 was only found in
PAUP* 4.0b10 (Swofford, 2001). Due to the large size variation in the
Caddo and Marshall counties, and mitotype A26b was observed in five
COI-COII region of Apis mellifera, the MP analysis was unrooted and
counties. Among the four mitotypes in Oklahoma, the average
no outgroup taxa were used. Gaps were treated as a missing
mitotype diversity, Hd was 0.59 (Table 3). For New Mexico, 47
character state for the maximum parsimony analysis. The reliability of
samples collected from 2005-8 were sequenced for the COI-COII
trees was tested with a bootstrap test (Felsenstein, 1985). Parsimony
marker. Mitotype A1a was the most common, followed by A1, A26a,
bootstrap analysis included 1,000 resamplings using the Branch and
A26b and A26. The average mitotype diversity, Hd was 0.62. Mitotype
Bound algorithm of PAUP* (Swofford, 2001).
A26 was the least common and was only found in Chaves county in
Table 2. Mitotype frequency and GenBank accession number. * 5’ and 3’ portion of sequence missing on GenBank. 1 = Collet et al. (2006). 2
= Franck et al. (2001). Mitotype
Amplicon size
Number
GenBank
A1
638
40
EF0336491
A1a
637, 577*
1
FJ4779842
A1d
638
92
FJ743639
A1e
638
7
GU326335
A4
830
2
EF0336501
A26
817
4
FJ4779902
A26a
830
16
FJ743640
A26b
831
4
FJ743641
A26c
830
1
FJ890929
A26d
830
1
GU326336
A29a
1006
3
FJ890930
A30
815
1
EF0336541
Total
172
Genetics of Africanized honey bees in the USA
183
Table 3. Summary statistics for mtDNA COI- COII polymorphisms in Africanized Apis mellifera mitotypes from the USA. *includes California, Arkansas and Florida; N = number of sequences; S = number of polymorphic sites; H = number of mitotypes; Hd = mitotype diversity; K = mean number of pairwise nucleotide differences; θg = theta per gene were the same; D+ and F+ statistics (Fu and Li 1993); Fu’s F’s statistic; Strobeck S statistic (Strobeck 1987); D = Tajima’s (1989) statistic; * p < 0.05; ** p < 0.02. State
N
S
H
Hd
K
θg
D+
F+
Fs
S
D
Oklahoma
55
197
4
0.59
46.14
1.104
0.886
1.202
1.211
0.446
1.368
New Mexico
47
198
5
0.62
66.66
1.184
-0.330
-0.49428
-0.659
0.901
1.558
Utah
20
199
5
0.69
65.80
1.462
0.512
0.568
0.463
0.744
0.603
Texas
44
295
6
0.39
50.65
1.889
-4.023**
-4.076**
1.412
0.376
-2.311*
Total*
172
396
12
0.51
54.90
1.378
-6.664**
-5.779**
0.952
0.454
-1.965*
2006. A total of 44 samples were sequenced from Texas, with
frequency and occurrence of ‘A’ lineage mitotypes in the USA. The
mitotype A1d being the most common, followed by A1 and A26a. The
level of genetic variation does indicate that COI-COII DNA sequences
average mitotype diversity for the Texas samples was 0.39. From the
could be used to detect the origin of AHB outbreaks in the USA for
20 samples subjected to DNA sequencing from Utah, a total of five
regulatory purposes.
mitotypes were observed. Mitotype A1e was the most common, and
Previous studies of COI-COII genetic variation of AHB in Mexico
this mitotype was unique to Utah, being observed in Washington and
(Kraus et al., 2007), Columbia (Prada et al., 2009), and Brazil and
Iron counties. The single Florida sample was mitotype A1d and the
Uruguay (Collet et al., 2006) has revealed several different mitotypes
three California samples were mitotype A29a.
of the ‘A’ lineage in each country. It is surprising that five of the nine
Fu and Li’s D+ and F+ statistics (1993), as well as Fu’s Fs statistic COI-COII AHB mitotypes observed in the USA have not been observed (Fu, 1997), and Tajima’s D statistic (1989) were computed on the
in Mexico and South America. This could be due to the fact that the
Oklahoma, New Mexico, Utah, Texas and all AHB samples (Table 3).
majority of the samples studied by Kraus et al. (2007), Prada et al.
These tests revealed that the Texas and total AHB population had an
(2009) and Collet et al. (2006) were subjected to PCR-restriction
excess of low frequency mitotypes, indicating population size
fragment length polymorphism (RFLP) analysis using the restriction
expansion. The New Mexico population had both negative and positive enzyme Dra I. PCR-RFLP is known to miss genetic variation that is values, while the Oklahoma population had positive values indicating
detected by DNA sequencing and based on our DNA sequences, Dra I
that balancing selection is occurring.
would not differentiate mitotypes A1d from A1 and A29a from A29.
For the maximum parsimony analysis, a total of 853 characters
Mitotypes A26abc are slightly different from mitotype A26 based on
were used, of which, 31 were parsimony informative (gaps treated as
Dra I restriction patterns and would fall in between mitotypes A25 and
missing). The MP analysis resulted in a single unrooted tree with a
A26 (Franck et al., 2001). This could be an explanation for the lack of
consistency index value of 0.881. The consensus MP phylogenetic tree mitotypes A1d, A26a, A26b, A26c, and A29a in other studies, and for had the representatives of the four A. mellifera lineages (M, C, O, and the high levels of intraspecific variation observed among ‘A’ lineage A) form distinct clades (Fig. 2). Within the ‘A’ lineage clade, mitoypes
mitotypes in the USA relative to other countries where the ‘A’ lineage
A26abc formed a sister group with an A26, A25 and A14 sequences
occurs.
from Namibia. Mitotypes A1, A1d, A1e, A30 and A4 formed a common
Kraus et al. (2007) using primarily Dra I PCR-RFLP data found two
clade with A27, A8 and A1abc mitotypes from Zambia and Brazil.
AHB mitotypes in central and southern Mexico, A1, A4. The A1
Finally, mitotypes A29a and A29 from Brazil formed a distinct clade
mitotype was more common than the A4 mitotype, which was also
relative to the other ‘A’ lineage mitotypes (Fig. 2).
observed in our samples from the USA (when combining mitotypes A1 and A1d). As previously mentioned, additional mitotypes probably exist in Mexico, but were not found due to the reduced sensitivity of
Discussion
PCR-RFLP to detect genetic variation compared with DNA sequencing. In Columbia, Prada et al. (2009) conducted an allozyme and mtDNA
The amount of COI-COII genetic variation of ‘A’ lineage honey bees in 16S and COI-COII Dra I analysis of 319 A. mellifera samples from 24 the USA is surprising for an invasive insect. Several of the mitotypes
locations. A total of eight ‘A’ lineage patterns were observed (A1, A4,
were widespread, i.e. A1 and A1d, while three of the mitotypes
A26, A28, A29, and A30), with A1 and A4 accounting for 83% of the
occurred only in Texas, A4 was only detected in Oklahoma, and A1e
samples. In Brazil and Uruguay Collet et al. (2006) also studied the
was only observed in Utah. Analysis of more samples, both
genetic structure of AHB using primarily Dra I PCR-RFLP. From 725
geographically and temporally, may provide more insight into the
colonies, a total of six AHB mitotypes were observed, with A4 being
184
Szalanski, Magnus
the most common followed by A1, A30, and then A26. The frequency
FERREIRA, K M; LINO E SILVA, O; ARIAS, M C; DEL LAMA, M A
of the A1 mitotypes increased towards the north of Brazil while the A4
(2009) Cytochrome-b variation in Apis mellifera samples and its
mitotype was more common in southern Brazil (Collet et al., 2006).
association with COI-COII patterns. Genética 135: 149-155.
This clinal variation, which also occurs in Africa (Collet et al., 2006)
FRANCK, P; GARNERY, L; LOISEAU, A; OLDROYD, B P; HEPBURN, H R;
may explain why A1 is more common in both Mexico (Kraus et al.,
SOLIGNAC, M; CORNUET, J-M (2001) Genetic diversity of the
2007) and the USA.
honey bee in Africa: microsatellite and mitochondrial data.
The origin of the A4 and A26 mitotypes can be attributed to the introduction of A. m. scutellata to Brazil in 1956 (Collet et al., 2006). Interestingly, Sheppard et al. (1999) using Dra I PCR-RFLP of honey bees from Argentina, attributed the A1 mitotype that he found to an African A. m. intermissa origin. The existence of more than one subspecies of African A. mellifera in the New World could explain the difference in migration of the different mitotypes, i.e. a greater proportion of A1 mitotypes in North America and more A4 mitotypes in South America. Future research is required to determine whether there is a difference in the levels of aggression amongst ‘A’ lineage mitotypes in
Heredity 86: 420-430. FU, X-Y; LI, W H (1993) Statistical tests of neutrality of mutations.
Genetics 133: 693-709. FU, X-Y (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: 915-925. GARNEY, L; SOLIGNAC, M; CELEBRANO, G; CORNUET, J M (1993) A simple test using restricted PCR-amplified mitochondrial DNA to study the genetic structure of Apis mellifera L. Experientia 49: 1016-1021. HALL, T A (1999) Bioedit: A user-friendly biological sequence
the USA, and whether there are any temporal or clinal patterns in the
alignment editor and analysis program for Windows 95/98/NT.
distribution of AHB in the USA.
Nucleic Acids Symposium Series 41: 95-98. HARTL, D L; CLARK, A G (1997) Principles of population genetics.
Acknowledgements
Sinauer Associates, Inc.; Sunderland, MA, USA. KRAUS, F B; FRANCK, P; VANDAME, R (2007) Asymmetric
We thank Ed Levi, Richard A Grantham, John Warner, Danielle
introgression of African genes in honey bee populations (Apis
Downey, and Lisa Bradley for providing samples. This research was
mellifera L.) in central Mexico. Heredity 99: 233-240.
supported in part by the University of Arkansas, Arkansas Agricultural Experiment Station.
NEI, M (1987) Molecular evolutionary genetics. Columbia University Press; New York, USA. PINTO, M A; JOHNSON, J S; RUBINK, W L; COULSON, R N;
References
PATTON, J C; SHEPPARD, W S (2003) Identification of Africanized
ANONYMOUS (2009) Map of the spread of Africanized honey bee by
of a rapid polymerase chain reaction-based assay. Annals of the
year. http://www.ars.usda.gov/research/docs.htm? docid=11059&page=6. ARIAS, M C; SHEPPARD, W S (1996) Molecular phylogenetics of honey
honey bee (Hymenoptera: Apidae) mitochondrial DNA: validation
Entomological Society of America 96: 679-684. PINTO, M A; RUBINK, W L; COULSON, R N; PATTON J C; JOHNSON J S (2004) Temporal pattern of Africanization in a feral
bee subspecies (Apis mellifera L.) inferred from mitochondrial DNA
honey bee population from Texas inferred from mitochondrial
sequences. Molecular Phylogenetics and Evolution 5: 557-566.
DNA. Evolution 58: 1047-1055.
CHAPMAN, N C; LIM, J; OLDROYD, B P (2008) Population genetics of
PINTO, M A; SHEPPARD, W S; JOHNSTON, J S; RUBINK, W L;
commercial and feral honey bees in western Australia. Journal of
COULSON, R N; SCHIFF, N M; KANDEMIR, I; PATTON, J C (2007)
Economic Entomology 101: 272-277.
Honey bees (Hymenoptera: Apidae) of African origin exist in non-
COLLET, T; FERREIRA, K P; ARIAS, M C; SOARES, A E E; DEL LAMA, M A (2006) Genetic structure of Africanized honey bee populations (Apis mellifera L.) from Brazil and Uruguay viewed through mitochondrial DNA COI-COII patterns. Heredity 97: 329-335. CROZIER, Y C; KOULIANOS, S; CROZIER, R H (1991) An improved test for Africanized honey bee mitochondrial DNA. Experientia 47: 968-969. FELSENSTEIN, J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791.
Afrincanized areas of the southern USA: Evidence from mitochondrial DNA. Annals of the Entomological Society of
America 100: 289-295. PRADA Q, C F; DURAN, J T; SALAMANCA G, G; DEL LAMA, M A (2009) Population genetics of Apis mellifera L. (Hymenoptera: Apidae) from Columbia. Journal of Apicultural Research 48: 3-10. DOI: 10.3896/IBRA.1.48.1.02 RINDERER, T E; BUCO, S M; RUBINK, W L; DALY, H V; STELZER, J A; RIGGIO, R M; BAPTISTA, F C (1993) Morphometric identification of Africanized and European honey bees using large reference populations. Apidologie 24: 569-585.
185
Genetics of Africanized honey bees in the USA
ROZAS, J; SANCHEZ-DELBARRIO, J-C; MESSEGUER, X; ROZAS, R (2003) DnaSP, DNA polymorphism analyzes by the coalescent and other methods. Bioinformatics 19: 2496-2497. RUTTNER, F (1988) Biogeography and taxonomy of honey bees. Springer-Verlag; Berlin Heidelberg, Germany. 284 pp. RUTTNER, F (1992) Naturgeschichte der Honigbienen, Ehrenwirth; Munich, Germany. SHEPPARD, W S; SMITH, D R (2000) Identification of African-derived bees in the Americas: a survey of methods. Annals of the
Entomological Society of America 93: 159-176. SHEPPARD, W S; RINDERER, T E; GARNEY, L; SHIMANUKI, H (1999) Analysis of Africanized honey bee mitochondrial DNA reveals
SUGDEN, D A; WILLIAMS, K R (1990) October 15: the day the bee arrived. Gleanings in Bee Culture 119: 18-21. SWOFFORD, D L (2001) PAUP: Phylogenetic analysis using parsimony
(and other methods). Version 4. Sinauer Associates; Sunderland, Massachussets, USA. SZALANSKI, A L; MCKERN, J A (2007) Multiplex PCR-RFLP diagnostics of the Africanized honey bee (Hymenoptera: Apidae). Sociobiology 50: 939-945. TAJIMA, F (1983) Evolutionary relationship of DNA sequences in finite populations. Genetics 105: 437-460. TAJIMA, F (1989) Statistical methods to test for nucleotide mutation hypothesis by DNA polymorphism. Genetics 123: 585-595.
further diversity of origin. Genetics and Molecular Biology 22: 73-75. THOMPSON, J D; HIGGINS, D G; GIBSON, T J (1994) CLUSTAL W: DOI: 10.1590/S1415-47571999000100015 SOLORZANO, C D; SZALANSKI, A L; KENCE, M; KENCE, A;
improving the sensitivity of progressive multiple sequence alignments through sequence weighting, position-specific gap
MCKERN, J A; AUSTIN, J W (2009) Phylogeography and
penalties and weight matrix choice. Nucleic Acids Research 22:
population genetics of honey bees (Apis mellifera L.) from Turkey
4673-4680.
based on COI-COII sequence data. Sociobiology 53: 237-246. STROBECK, C (1987) Average number of nucleotide difference in a sample from a single subpopulation: a test for population subdivision. Genetics 117: 149-153.
© IBRA 2010
DOI 10.3896/IBRA.1.49.2.06
ORIGINAL RESEARCH ARTICLE
Mitochondrial DNA characterization of Africanized honey bee (Apis mellifera L.) populations from the USA Allen L Szalanski1* and Roxane M Magnus1 1
Social Insects Genetics Lab, Department of Entomology, University of Arkansas, Fayetteville, AR 72701, USA.
Received 15 April 2009, accepted subject to revision 9 December 2009, accepted for publication 23 February 2010. *Corresponding author: Email: [email protected]
Summary We carried out a study which involved DNA sequencing of a portion of the mitochondrial DNA COI-COII region of Africanized honey bees (AHB) from the USA. A total of 12 mitotypes were observed, of which seven have not been previously described. Of the 172 samples, two mitotypes, A1 and A1d, accounted for 77% of the observed mitotypes, while mitotypes A1a, A26c, A26d, A29a, and A30 were only observed once. A possible reason why these new mitotypes have not been described before is because previous studies on AHB in the new world have relied primarily on PCR-restriction fragment length polymorphism (RFLP), which is less sensitive than DNA sequence data. Multiple mitotypes of ‘A’ lineage honey bees have previously been observed in South America and Mexico using PCR-RFLP and DNA sequence analysis. Our findings are consistent with previous studies of AHB genetic variation from central Mexico, Columbia, and northern Brazil, in that the A1 mitotype was more common than the A4. Maximum parsimony analysis revealed that all of the ‘A’ lineage Apis mellifera mitotypes formed a distinct clade relative to representatives of the ‘M’, ‘C’, and ‘O’ lineages. Statistical analysis of the mitotype frequencies in the USA revealed an excess of low frequency mitotypes, indicating that the population size is expanding. The amount of genetic variation observed in Africanized honey bees in the USA therefore supports the idea that there have been multiple introductions of AHB into the country.
Caracterización mitocondrial de poblaciones de abeja (Apis
mellifera L.) africanizada de los EEUU Resumen Se realizó un estudio mediante la secuenciación de una porción de la región COI-COII del ADN mitocondrial en las abejas de la miel africanizadas de los EEUU. Se observaron un total de 12 mitotipos, de los cuales siete no se habían descrito anteriormente. De las 172 muestras, dos mitotipos, A1 y A1d, representaron el 77% de los mitotipos detectados, mientras que los mitotipos A1a, A26c, A26d, A29a y A30 sólo fueron observados una vez. Una posible razón de por qué estos nuevos mitotipos no se hayan descrito antes es porque los estudios previos sobre la abeja africanizada en el nuevo mundo se habían basado principalmente en PCR-Polimorfismo de longitud de los fragmentos de restricción (RFLP), que es menos sensible que los datos de la secuencia de ADN. Varios mitotipos en las abejas de miel del linaje “A” habían sido previamente observados en América del Sur y México, utilizando PCR-RFLP y el análisis de secuencias de ADN. Nuestros resultados son consistentes con los estudios previos sobre la variación genética de la abeja africanizada en México central, Colombia y el norte de Brasil, en los que el mitotipo A1 es más común que el A4. El análisis de máxima parsimonia reveló que todos los mitotipos del linaje “A” de
Apis mellifera formaron un clado distinto en relación con los representantes de los linajes “O”, “M” y “C”. El análisis estadístico de las frecuencias de los mitotipos en los EEUU reveló un exceso de mitotipos de baja frecuencia, lo que indica que el tamaño de la población está en expansión. La cantidad de la variación genética observada en las abejas de miel africanizadas de los EEUU por tanto, apoya la idea de que ha habido introducciones múltiples de abejas africanizadas en el país.
Keywords: Apis mellifera, COI-COII intergenic region, Africanized honey bee, mtDNA, USA
Introduction
distributed across the world due to multiple migrations and introductions
Apis mellifera L. is native to Europe, Africa, and Asia (including Saudi
that are present in different regions of the world. These subspecies
Arabia, Iran and the Ural mountains of Russia). It is currently widely
have been classified into five main lineages: ‘C’ (the Carnica
(Ruttner, 1988). Apis mellifera includes about two dozen subspecies
178
Szalanski, Magnus
group that includes A. m. carnica and A. m. ligustica); ‘M’ (the north
al., 2003; 2004; 2007). In contrast to cyt b, the mtDNA cytochome
and western European honey bees that include A. m. mellifera,
oxidase I (COI) and COII genes for A. mellifera exhibit a high degree
A.m. iberica, and A. m. intermissa); ‘A’ (the African group that
of genetic variation within and amongst lineages, and is the preferred
includes A. m. scutellata, A. m. capensis, A. m. lamarckii, A. m.
molecular marker for detecting intraspecific genetic variation. This
litorea, A. m. adansonii, and A. m. unicolor); ‘Y’ (Ethiopia, Franck et
marker has been used for genetic analysis of honey bee populations
al., 2001); and the ‘O’ group (the Oriental or Middle Eastern group
from: Turkey (Solorzano et al., 2009); Mexico (Kraus et al., 2007);
which includes A. m. anatolica, A. m. caucasica, A. m. syriaca,
South America (Collet et al., 2006; Ferreira et al., 2009; Prada et al.,
A. m. pomonella, and A. m. cypria) (Ruttner, 1992). At the molecular
2009); Africa (Franck et al., 2001); and Australia (Chapman et al.,
level, these lineages are genetically divergent based on nuclear and
2008). Previous studies of COI-COII genetic variation of AHB in
mitochondrial DNA markers (Arias and Sheppard, 1996; Franck et al.,
Mexico, Columbia, Brazil and Uruguay have revealed several different
2001).
mitotypes of the ‘A’ lineage in each country, with two mitotypes
The Africanized honey bee (AHB) was first detected in Texas in
occurring in central Mexico (Kraus et al., 2007), six in Columbia
1990 (Sugden and Williams, 1990), and by 2009 had spread to a total (Prada et al., 2009), and five in Brazil and Uruguay (Collet et al., of ten states (Anonymous, 2009). The hybrid in the USA is virtually
2006). To date there is no information on COI-COII DNA sequence
indistinguishable in the field from the European honey bee (EHB), and variation of AHB in the USA, so the objective of our study was to requires a morphometric analysis for identification (Rinderer et al.,
determine the genetic diversity of AHB from the USA based on COI-
1993). Mitochondrial DNA (mtDNA) is an ideal genetic marker for
COII DNA sequence data.
identifying AHB, since a single worker can represent the entire colony (Sheppard and Smith, 2000). Introgression of AHB genes using a mtDNA marker is, however, not detectable if an EHB queen has mated with AHB drones.
Materials and methods Sampling
Previous studies of the molecular genetics of AHB in the USA were Specimens of adult worker honey bees were collected from Utah, New focused on molecular diagnostics, and primarily used a region of the
Mexico, Oklahoma, Texas, California, Florida and Arkansas and stored
mitochondrial DNA (mtDNA) genome, cytochrome b (cyt b) which has
in 70-100% ethanol until processed for DNA extraction (Table 1, Fig. 1).
a relatively low level of intraspecific variation in A. mellifera (Crozier et Samples were collected from managed colonies, feral colonies and
al., 1991), making it ideal for molecular diagnostic purposes (Pinto et
swarms. The two Arkansas samples were from two counties in the
Fig. 1. Frequency and locations of ‘A’ lineage Apis mellifera mitotypes from seven states.
Genetics of Africanized honey bees in the USA
179
Table 1. Distribution and frequency of the studied Apis mellifera ‘A’ lineage mitotypes from the USA. State
County
Lat/Long
Mitotype (n)
Arkansas (2)
Lafayette
33.26, -93.59
A1d(1)
Miller
33.32, -93.87
A1(1)
California (3)
San Diego
33.02, -116.77
A26a(1), A29a(2)
New Mexico (47)
Chaves
33.36, -104.47
A1(1), A1d(3), A26(2)
Curry
34.57, -103.35
A1(1), A1d(4), A26b(1)
De Baca
34.47, -104.24
A1(1)
Doña Ana
32.31, -106.77
A1(4), A1d(4), A26a(1)
Eddy
32.47, -104.3
A26a(1)
Grant
32.73, -108.38
A1d(1)
Guadalupe
34.86, -104.78
A1d(2)
Lincoln
33.74, -105.46
A1d(4), A26a(1)
Otero
32.62, -105.73
A1(1), A1d(4)
Roosevelt
34.02, -103.48
A1(1), A1d(4), A26a(1), A26b(3)
Santa Fe
35.51, -105.98
A1d(1)
Socorro
34.02, -106.93
A1(1)
Beckham
35.26, -99.69
A1d(1)
Blaine
35.88, -98.43
A1d(2)
Bryan
33.97, -96.25
A1(1)
Caddo
35.18, -98.38
A1(3), A4(1), A26a(1)
Carter
34.25, -97.29
A1d(1)
Cleveland
35.2, -97.33
A1(1), A1d(1)
Coal
34.6, -96.3
A1d(2)
Comanche
34.66, -98.46
A1(2), A1d(1)
Cotton
34.28, -98.37
A1(1), A1d(1)
Custer
35.64, -99.01
A1d(1), A26a(1)
Dewey
35.99, -99.00
A1d(3)
Grady
35.02, -97.89
A1(1)
Greer
34.93, -99.56
A1(1)
Harmon
34.74, -99.84
A1(1), A1d(1)
Hughes
35.04, -96.26
A1d(1)
Jefferson
34.10, -97.84
A1d(1)
Kiowa
34.92, -98.98
A1(1), A1d(7)
Love
33.95, -97.25
A1d(1), A26a(1)
Marshall
34.03, -96.77
A4(1), A26a(1)
McClain
35.00, -97.44
A1d(1), A26a(1)
McCurtain
34.11, -94.77
A1(1), A1d(1)
Murray
34.49, -97.07
A1(1)
Oklahoma
35.48, -97.53
A1(1), A26a(1)
Payne
35.48, -97.53
A1d(1)
Pottawatomie
35.20, -96.94
A1d(1)
Oklahoma (55)
180
Szalanski, Magnus
Table 1. Cont’d Distribution and frequency of the studied Apis mellifera ‘A’ lineage mitotypes from the USA. State
County
Lat/Long
Mitotype (n)
Stephens
34.48, -97.86
A1(1), A1d(1)
Tillman
34.38, -98.92
A1(1)
Washita
35.29, -98.99
A1d(1)
Washington
37.28, -113.52
A1(2), A1d(3), A1e(2), A26(1), A26a(1)
Kane
37.29, -111.89
A1d(2)
Iron
37.86, -113.28
A1(2), A1e(5), A26a(2)
Florida (1)
Broward
26.12, -80.24
A1d(1)
Texas (44)
Armstrong
34.97, -101.35
A1d(2)
Bastrop
30.10, -97.31
A1d(1)
Bexar
29.45, -98.52
A1(1)
Bosque
31.9, -97.63
A1d(1)
Calhoun
28.44, -96.61
A1(1)
Cameron
26.15, -97.45
A1(2)
Clay
33.79, -98.21
A1d(1)
Dawson
32.74, -101.95
A1d(1)
Dimmit
28.42, -99.75
A1d(1)
Fisher
32.74, -100.4
A1d(1)
Gaines
32.74, -102.64
A1d(1)
Gray
35.41, -100.81
A1d(2)
Gregg
32.48, -94.81
A1(1)
Guadalupe
29.58, -97.95
A29a(1)
Hamilton
31.70, -98.11
A1d(1)
Hidalgo
26.40, -98.18
A1(1)
Johnson
32.38, -97.36
A1d(1)
Jones
32.74, -99.88
A26a(1)
Kerr
30.06, -99.35
A1d(3)
Lee
30.31, -96.96
A1(1), A1a(1)
Lubbock
33.61, -101.82
A26(1), A30(1)
Matagorda
28.78, -96.00
A26c(1)
Pecos
30.78, -102.72
A1d(1)
Randall
34.97, -101.90
A1d(4)
Refugio
28.32, -97.17
A1d(1)
Swisher
34.53, -101.73
A1d(1)
Taylor
32.31, -99.88
A1d(1), A26a(1)
Travis
30.33, -97.78
A1d(1)
Uvalde
29.35, -99.76
A1d(1)
Victoria
28.8, -96.97
A1(1), A1d(1)
Williamson
30.32, -97.62
A1d(1)
Young
33.18, -98.7
A1(1)
Utah (20)
Genetics of Africanized honey bees in the USA
181
extreme south west corner of the state, and were collected by the
Mitochondrial analysis
Arkansas Plant Board in 2005 as part of their AHB monitoring
Genomic DNA from individual honey bee thoraces was extracted using
programme. The Oklahoma samples were collected in 2004-8
the Qiagen DNeasy extraction kit (QIAGEN; Valencia, CA, USA)
primarily from feral colonies and swarms as part of the AHB
according to the manufacturer’s protocol and per Solorzano et al.,
diagnostics programme conducted by Oklahoma State University.
(2009). A 637 to 1006 bp region of the COI-COII intergenic region
Identification of samples collected from New Mexico, Oklahoma and
was PCR amplified in a Techne T-412 thermal cycler (Techne Inc;
Arkansas from 2004-6 and from the Utah, Florida and California
Burlington, NJ, USA) using primers E2 and H2 (Garney et al., 1993).
samples was done using multiplex PCR following Szalanski and
PCR was conducted using a profile consisting of an initial denaturation
McKern (2007). Additional samples collected during 2007 and 2008
of 94oC for 2 min followed by 35 cycles of 94oC for 45s, 46oC for 45s,
from Oklahoma and New Mexico were determined as AHB using PCR-
and 72oC for 45s, and then a final extension of 72oC for 5 min.
RFLP (Pinto et al., 2003) by R A Grantham (Oklahoma State
Amplicons were separated using 1% agarose gel electrophoresis and
University, USA). Samples collected from Texas from 1991 to 2008
photo documented using a BioDoc-It™ Imaging System (UVP, Inc.;
were identified as Africanized (n = 26), Africanized with evidence of
Upland, CA, USA) per Solorzano et al., (2009). PCR products were
introgression of European genes (n = 11) or European with evidence
purified using Microcon-PCR Filter Units (Millipore; Bedford, MA, USA),
of introgression of Africanized genes (n = 6) using FABIS (Rinderer et
and sent to the University of Arkansas Medical Sciences DNA
al., 1993) by L Bradley (Texas A&M University, USA). Utah samples
Sequencing Core Facility (Little Rock, AR, USA) for direct sequencing
were collected in 2008 and 2009 by the Utah Department of
in both directions. DNA sequences new to this study were deposited
Agriculture and Food. Voucher specimens are deposited at the
in NCBI GenBank (http://www.ncbi.nlm.nih.gov/) as accession
Arthropod Museum, Department of Entomology, University of
numbers FJ743632 to FJ743641, FJ890929, FJ890930, GU326335 and
Arkansas, Fayetteville, AR, USA.
GU326336.
Fig. 2. Phylogenetic relationship among mitotypes representing four lineages of Apis mellifera based on an unrooted maximum parsimony heuristic analysis. Maximum parsimony bootstrap values (>50%) are provided, and GenBank accession numbers are provided for each mitotype.
182
Szalanski, Magnus
Genetic diversity analysis
Results
DNA sequences were aligned with CLUSTAL W (Thompson et al., 1994) using Bioedit v5.0.7 (Hall, 1999). Designation of mitotypes was done
Table 1 shows the DNA sequencing analysis of honey bee samples
by comparing sequences with those available on GenBank. Number of
collected from New Mexico, Utah, Texas, Florida, Oklahoma,
mitotypes and their frequencies were determined both visually and
California, and Arkansas. A total of 12 mitotypes were observed,
with the program DNAsp version 4.10.9 (Rozas et al., 2003). DNAsp
which ranged from 637 bp to 831 bp in size (Table 2). Four mitotypes
was also used to estimate the following variables: haplotypic diversity
have previously been described. Mitotype A4 was identical to GenBank
(Hd) (Nei, 1987), and Nei’s Nm value. Nucleotide diversity was
EF033650 from Brazil, mitotype A1 to EF033649 from Brazil, mitotype
interpreted as the average proportion of nucleotide differences
A30 to EF033654 from Brazil, and mitotype A26 to FJ477990 from
between all possible pairs of sequences in the sample (Hartl and
Namibia. Based on a BLAST search of GenBank DNA sequences, five
Clark, 1997), mean number of pairwise nucleotide differences (K)
of the 12 mitotypes observed have not previously been described.
equation A3 (Tajima, 1983), number of polymorphic sites (S), and the Mitotypes A1d and A1e were most similar to EF033649 mitotype A1 parameter Өg. The parameter θ is the proportion of nucleotide sites
(0.5% divergence), mitotype A29a was closest to EF033653 mitotype
that are expected to be polymorphic in any suitable sample from this
A29 (0.3% divergence) and mitotypes A26a to A26d were most
region of the genome (Hartl and Clark, 1997). To test for neutral
similar to FJ477990 mitotype A26 (1.6 to 1.8% divergence). Overall,
mutation, Tajima’s D (Tajima, 1989), and D* and F* (Fu and Li, 1993) mitotypes A1 and A1d were the most common, accounting for 77% of were calculated. To examine demographic stability, Fu’s Fs statistic
samples, while the rest of the 10 observed mitotypes accounted for
(Fu, 1997) (based on mitotype distribution) was used.
the remaining 23% of the samples.
For the phylogenetic analysis, DNA sequences were aligned using
In Arkansas, two AHB samples were found, one A1 and one A1d
CLUSTAL W (Thompson et al., 1994) using DNA sequences from this
mitotypes (Table 1, Fig. 1). In Oklahoma, a total of 55 samples from
study and additional ones from GenBank (Fig. 2). Maximum
28 counties were sequenced, with mitotype A1d being the most
parsimony (MP) analysis on the alignments was conducted using
common, followed by A1, A26a and A4. Mitotype A4 was only found in
PAUP* 4.0b10 (Swofford, 2001). Due to the large size variation in the
Caddo and Marshall counties, and mitotype A26b was observed in five
COI-COII region of Apis mellifera, the MP analysis was unrooted and
counties. Among the four mitotypes in Oklahoma, the average
no outgroup taxa were used. Gaps were treated as a missing
mitotype diversity, Hd was 0.59 (Table 3). For New Mexico, 47
character state for the maximum parsimony analysis. The reliability of
samples collected from 2005-8 were sequenced for the COI-COII
trees was tested with a bootstrap test (Felsenstein, 1985). Parsimony
marker. Mitotype A1a was the most common, followed by A1, A26a,
bootstrap analysis included 1,000 resamplings using the Branch and
A26b and A26. The average mitotype diversity, Hd was 0.62. Mitotype
Bound algorithm of PAUP* (Swofford, 2001).
A26 was the least common and was only found in Chaves county in
Table 2. Mitotype frequency and GenBank accession number. * 5’ and 3’ portion of sequence missing on GenBank. 1 = Collet et al. (2006). 2
= Franck et al. (2001). Mitotype
Amplicon size
Number
GenBank
A1
638
40
EF0336491
A1a
637, 577*
1
FJ4779842
A1d
638
92
FJ743639
A1e
638
7
GU326335
A4
830
2
EF0336501
A26
817
4
FJ4779902
A26a
830
16
FJ743640
A26b
831
4
FJ743641
A26c
830
1
FJ890929
A26d
830
1
GU326336
A29a
1006
3
FJ890930
A30
815
1
EF0336541
Total
172
Genetics of Africanized honey bees in the USA
183
Table 3. Summary statistics for mtDNA COI- COII polymorphisms in Africanized Apis mellifera mitotypes from the USA. *includes California, Arkansas and Florida; N = number of sequences; S = number of polymorphic sites; H = number of mitotypes; Hd = mitotype diversity; K = mean number of pairwise nucleotide differences; θg = theta per gene were the same; D+ and F+ statistics (Fu and Li 1993); Fu’s F’s statistic; Strobeck S statistic (Strobeck 1987); D = Tajima’s (1989) statistic; * p < 0.05; ** p < 0.02. State
N
S
H
Hd
K
θg
D+
F+
Fs
S
D
Oklahoma
55
197
4
0.59
46.14
1.104
0.886
1.202
1.211
0.446
1.368
New Mexico
47
198
5
0.62
66.66
1.184
-0.330
-0.49428
-0.659
0.901
1.558
Utah
20
199
5
0.69
65.80
1.462
0.512
0.568
0.463
0.744
0.603
Texas
44
295
6
0.39
50.65
1.889
-4.023**
-4.076**
1.412
0.376
-2.311*
Total*
172
396
12
0.51
54.90
1.378
-6.664**
-5.779**
0.952
0.454
-1.965*
2006. A total of 44 samples were sequenced from Texas, with
frequency and occurrence of ‘A’ lineage mitotypes in the USA. The
mitotype A1d being the most common, followed by A1 and A26a. The
level of genetic variation does indicate that COI-COII DNA sequences
average mitotype diversity for the Texas samples was 0.39. From the
could be used to detect the origin of AHB outbreaks in the USA for
20 samples subjected to DNA sequencing from Utah, a total of five
regulatory purposes.
mitotypes were observed. Mitotype A1e was the most common, and
Previous studies of COI-COII genetic variation of AHB in Mexico
this mitotype was unique to Utah, being observed in Washington and
(Kraus et al., 2007), Columbia (Prada et al., 2009), and Brazil and
Iron counties. The single Florida sample was mitotype A1d and the
Uruguay (Collet et al., 2006) has revealed several different mitotypes
three California samples were mitotype A29a.
of the ‘A’ lineage in each country. It is surprising that five of the nine
Fu and Li’s D+ and F+ statistics (1993), as well as Fu’s Fs statistic COI-COII AHB mitotypes observed in the USA have not been observed (Fu, 1997), and Tajima’s D statistic (1989) were computed on the
in Mexico and South America. This could be due to the fact that the
Oklahoma, New Mexico, Utah, Texas and all AHB samples (Table 3).
majority of the samples studied by Kraus et al. (2007), Prada et al.
These tests revealed that the Texas and total AHB population had an
(2009) and Collet et al. (2006) were subjected to PCR-restriction
excess of low frequency mitotypes, indicating population size
fragment length polymorphism (RFLP) analysis using the restriction
expansion. The New Mexico population had both negative and positive enzyme Dra I. PCR-RFLP is known to miss genetic variation that is values, while the Oklahoma population had positive values indicating
detected by DNA sequencing and based on our DNA sequences, Dra I
that balancing selection is occurring.
would not differentiate mitotypes A1d from A1 and A29a from A29.
For the maximum parsimony analysis, a total of 853 characters
Mitotypes A26abc are slightly different from mitotype A26 based on
were used, of which, 31 were parsimony informative (gaps treated as
Dra I restriction patterns and would fall in between mitotypes A25 and
missing). The MP analysis resulted in a single unrooted tree with a
A26 (Franck et al., 2001). This could be an explanation for the lack of
consistency index value of 0.881. The consensus MP phylogenetic tree mitotypes A1d, A26a, A26b, A26c, and A29a in other studies, and for had the representatives of the four A. mellifera lineages (M, C, O, and the high levels of intraspecific variation observed among ‘A’ lineage A) form distinct clades (Fig. 2). Within the ‘A’ lineage clade, mitoypes
mitotypes in the USA relative to other countries where the ‘A’ lineage
A26abc formed a sister group with an A26, A25 and A14 sequences
occurs.
from Namibia. Mitotypes A1, A1d, A1e, A30 and A4 formed a common
Kraus et al. (2007) using primarily Dra I PCR-RFLP data found two
clade with A27, A8 and A1abc mitotypes from Zambia and Brazil.
AHB mitotypes in central and southern Mexico, A1, A4. The A1
Finally, mitotypes A29a and A29 from Brazil formed a distinct clade
mitotype was more common than the A4 mitotype, which was also
relative to the other ‘A’ lineage mitotypes (Fig. 2).
observed in our samples from the USA (when combining mitotypes A1 and A1d). As previously mentioned, additional mitotypes probably exist in Mexico, but were not found due to the reduced sensitivity of
Discussion
PCR-RFLP to detect genetic variation compared with DNA sequencing. In Columbia, Prada et al. (2009) conducted an allozyme and mtDNA
The amount of COI-COII genetic variation of ‘A’ lineage honey bees in 16S and COI-COII Dra I analysis of 319 A. mellifera samples from 24 the USA is surprising for an invasive insect. Several of the mitotypes
locations. A total of eight ‘A’ lineage patterns were observed (A1, A4,
were widespread, i.e. A1 and A1d, while three of the mitotypes
A26, A28, A29, and A30), with A1 and A4 accounting for 83% of the
occurred only in Texas, A4 was only detected in Oklahoma, and A1e
samples. In Brazil and Uruguay Collet et al. (2006) also studied the
was only observed in Utah. Analysis of more samples, both
genetic structure of AHB using primarily Dra I PCR-RFLP. From 725
geographically and temporally, may provide more insight into the
colonies, a total of six AHB mitotypes were observed, with A4 being
184
Szalanski, Magnus
the most common followed by A1, A30, and then A26. The frequency
FERREIRA, K M; LINO E SILVA, O; ARIAS, M C; DEL LAMA, M A
of the A1 mitotypes increased towards the north of Brazil while the A4
(2009) Cytochrome-b variation in Apis mellifera samples and its
mitotype was more common in southern Brazil (Collet et al., 2006).
association with COI-COII patterns. Genética 135: 149-155.
This clinal variation, which also occurs in Africa (Collet et al., 2006)
FRANCK, P; GARNERY, L; LOISEAU, A; OLDROYD, B P; HEPBURN, H R;
may explain why A1 is more common in both Mexico (Kraus et al.,
SOLIGNAC, M; CORNUET, J-M (2001) Genetic diversity of the
2007) and the USA.
honey bee in Africa: microsatellite and mitochondrial data.
The origin of the A4 and A26 mitotypes can be attributed to the introduction of A. m. scutellata to Brazil in 1956 (Collet et al., 2006). Interestingly, Sheppard et al. (1999) using Dra I PCR-RFLP of honey bees from Argentina, attributed the A1 mitotype that he found to an African A. m. intermissa origin. The existence of more than one subspecies of African A. mellifera in the New World could explain the difference in migration of the different mitotypes, i.e. a greater proportion of A1 mitotypes in North America and more A4 mitotypes in South America. Future research is required to determine whether there is a difference in the levels of aggression amongst ‘A’ lineage mitotypes in
Heredity 86: 420-430. FU, X-Y; LI, W H (1993) Statistical tests of neutrality of mutations.
Genetics 133: 693-709. FU, X-Y (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147: 915-925. GARNEY, L; SOLIGNAC, M; CELEBRANO, G; CORNUET, J M (1993) A simple test using restricted PCR-amplified mitochondrial DNA to study the genetic structure of Apis mellifera L. Experientia 49: 1016-1021. HALL, T A (1999) Bioedit: A user-friendly biological sequence
the USA, and whether there are any temporal or clinal patterns in the
alignment editor and analysis program for Windows 95/98/NT.
distribution of AHB in the USA.
Nucleic Acids Symposium Series 41: 95-98. HARTL, D L; CLARK, A G (1997) Principles of population genetics.
Acknowledgements
Sinauer Associates, Inc.; Sunderland, MA, USA. KRAUS, F B; FRANCK, P; VANDAME, R (2007) Asymmetric
We thank Ed Levi, Richard A Grantham, John Warner, Danielle
introgression of African genes in honey bee populations (Apis
Downey, and Lisa Bradley for providing samples. This research was
mellifera L.) in central Mexico. Heredity 99: 233-240.
supported in part by the University of Arkansas, Arkansas Agricultural Experiment Station.
NEI, M (1987) Molecular evolutionary genetics. Columbia University Press; New York, USA. PINTO, M A; JOHNSON, J S; RUBINK, W L; COULSON, R N;
References
PATTON, J C; SHEPPARD, W S (2003) Identification of Africanized
ANONYMOUS (2009) Map of the spread of Africanized honey bee by
of a rapid polymerase chain reaction-based assay. Annals of the
year. http://www.ars.usda.gov/research/docs.htm? docid=11059&page=6. ARIAS, M C; SHEPPARD, W S (1996) Molecular phylogenetics of honey
honey bee (Hymenoptera: Apidae) mitochondrial DNA: validation
Entomological Society of America 96: 679-684. PINTO, M A; RUBINK, W L; COULSON, R N; PATTON J C; JOHNSON J S (2004) Temporal pattern of Africanization in a feral
bee subspecies (Apis mellifera L.) inferred from mitochondrial DNA
honey bee population from Texas inferred from mitochondrial
sequences. Molecular Phylogenetics and Evolution 5: 557-566.
DNA. Evolution 58: 1047-1055.
CHAPMAN, N C; LIM, J; OLDROYD, B P (2008) Population genetics of
PINTO, M A; SHEPPARD, W S; JOHNSTON, J S; RUBINK, W L;
commercial and feral honey bees in western Australia. Journal of
COULSON, R N; SCHIFF, N M; KANDEMIR, I; PATTON, J C (2007)
Economic Entomology 101: 272-277.
Honey bees (Hymenoptera: Apidae) of African origin exist in non-
COLLET, T; FERREIRA, K P; ARIAS, M C; SOARES, A E E; DEL LAMA, M A (2006) Genetic structure of Africanized honey bee populations (Apis mellifera L.) from Brazil and Uruguay viewed through mitochondrial DNA COI-COII patterns. Heredity 97: 329-335. CROZIER, Y C; KOULIANOS, S; CROZIER, R H (1991) An improved test for Africanized honey bee mitochondrial DNA. Experientia 47: 968-969. FELSENSTEIN, J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791.
Afrincanized areas of the southern USA: Evidence from mitochondrial DNA. Annals of the Entomological Society of
America 100: 289-295. PRADA Q, C F; DURAN, J T; SALAMANCA G, G; DEL LAMA, M A (2009) Population genetics of Apis mellifera L. (Hymenoptera: Apidae) from Columbia. Journal of Apicultural Research 48: 3-10. DOI: 10.3896/IBRA.1.48.1.02 RINDERER, T E; BUCO, S M; RUBINK, W L; DALY, H V; STELZER, J A; RIGGIO, R M; BAPTISTA, F C (1993) Morphometric identification of Africanized and European honey bees using large reference populations. Apidologie 24: 569-585.
185
Genetics of Africanized honey bees in the USA
ROZAS, J; SANCHEZ-DELBARRIO, J-C; MESSEGUER, X; ROZAS, R (2003) DnaSP, DNA polymorphism analyzes by the coalescent and other methods. Bioinformatics 19: 2496-2497. RUTTNER, F (1988) Biogeography and taxonomy of honey bees. Springer-Verlag; Berlin Heidelberg, Germany. 284 pp. RUTTNER, F (1992) Naturgeschichte der Honigbienen, Ehrenwirth; Munich, Germany. SHEPPARD, W S; SMITH, D R (2000) Identification of African-derived bees in the Americas: a survey of methods. Annals of the
Entomological Society of America 93: 159-176. SHEPPARD, W S; RINDERER, T E; GARNEY, L; SHIMANUKI, H (1999) Analysis of Africanized honey bee mitochondrial DNA reveals
SUGDEN, D A; WILLIAMS, K R (1990) October 15: the day the bee arrived. Gleanings in Bee Culture 119: 18-21. SWOFFORD, D L (2001) PAUP: Phylogenetic analysis using parsimony
(and other methods). Version 4. Sinauer Associates; Sunderland, Massachussets, USA. SZALANSKI, A L; MCKERN, J A (2007) Multiplex PCR-RFLP diagnostics of the Africanized honey bee (Hymenoptera: Apidae). Sociobiology 50: 939-945. TAJIMA, F (1983) Evolutionary relationship of DNA sequences in finite populations. Genetics 105: 437-460. TAJIMA, F (1989) Statistical methods to test for nucleotide mutation hypothesis by DNA polymorphism. Genetics 123: 585-595.
further diversity of origin. Genetics and Molecular Biology 22: 73-75. THOMPSON, J D; HIGGINS, D G; GIBSON, T J (1994) CLUSTAL W: DOI: 10.1590/S1415-47571999000100015 SOLORZANO, C D; SZALANSKI, A L; KENCE, M; KENCE, A;
improving the sensitivity of progressive multiple sequence alignments through sequence weighting, position-specific gap
MCKERN, J A; AUSTIN, J W (2009) Phylogeography and
penalties and weight matrix choice. Nucleic Acids Research 22:
population genetics of honey bees (Apis mellifera L.) from Turkey
4673-4680.
based on COI-COII sequence data. Sociobiology 53: 237-246. STROBECK, C (1987) Average number of nucleotide difference in a sample from a single subpopulation: a test for population subdivision. Genetics 117: 149-153.
