Both the great divergence among Canary Islands nucleomorphs (6 = 0.02 1) compared with the maximum nucleomorph distance in all other populations (6.
Mitochondrial DNA Variation and Genetic Structure in Old-World Populations of Drosophila subobscura’ Juan M. Afonso,* Andrea Volz,“fMariano Hernandez,* Henriette Ruttkay,? AI. Gonzalez,* Jose M. Larruga,” Vicente M. Cabrera,* and Diether Sperlicht *Department o f Genetics, University of La Laguna, Canary Islands; and TLehrstuhl fur Populationsgenetik, Universitat Tubingen
To discover the relation between mitochondrial DNA (mtDNA) polymorphism and the geographic population structure of Drosophila subobscura previously established for other genetic traits, a wide Paleartic survey was carried out. A total of 24 nucleomorphs was observed among 26 1 isofemale lines assayed by 11 restriction endonucleases with 38 different sites in the mtDNA cleavage map. The differentiation of the Canary Islands populations (6 = 0.0119) compared with the mean among all the other continental and insular populations (6 = 0.0002) is striking. Both the great divergence among Canary Islands nucleomorphs (6 = 0.02 1) compared with the maximum nucleomorph distance in all other populations (6 = 0.0 17) and the abundance of endemic nucleomorphs (11) on the Canary Islands (50% of the total number of different nucleomorphs found in the entire distribution area) suggest that this molecular differentiation most probably results from the very old age of the Canary Islands populations rather than from drift and founder effects. Introduction
Because of its relative abundance, its extensive geographic distribution, and its rich chromosomal polymorphism, Drosophila subobscura has become the most intensively investigated European species of the genus Drosophila with regard to population genetics. There is evidence, at different levels, that natural populations of D. subobscura are geographically structured. Differences have been observed between northern marginal and southern central populations, with respect to average chromosomal heterozygosity, viability distribution, and lethal gene frequencies (Sperlich et al. 1977, 1980; Pfriem and Sperlich 1982). On the other hand, no differences were found normally for enzymatic variation (Saura et al. 1973). The existence of strong north-south clines for quantitative variation, such as wing size (Prevosti 1955; Pfriem 1983), gene arrangement frequencies (Prevosti 1964; Krimbas 1965; Krimbas and Loukas 1980), and exceptionally for allozyme frequencies (Pinsker and Sperlich 1979; Prevosti et al. 1983), is evident. Isolated insular populations of this species seem to possess generally primitive features in their chromosomal polymorphism (Prevosti et al. 1975), but with respect to allozyme variation the populations from the Canary Islands are clearly distinguished from both continental populations and other insular populations (Larruga et al. 1983). There is also evidence that D. subobscura populations 1. Key words: Drosophilasubobscuru,mitochondrial DNA, population structure, restriction-site polymorphisms. Address for correspondence and reprints: Dr. D. Sperlich, Lehrstuhl ftir Populationsgenetik, Fakultit ftir Biologie, Universtit Tiibingen, Auf der Morgenstelle 28, D-7400 Tiibingen, Federal Republic of Germany. Mol. Biol. Evol. 7(2):123-142.
1990. 0 1990 by The University of Chicago. All rights reserved. 0737-4038/90/0702-0002$02.00
123
124 Afonso et al.
are structured on a fine scale, both temporally and spatially. Seasonal changes have been found for gene arrangement frequencies (De Frutos and Prevosti 1984). Nonrandom microdispersion of inversion polymorphism has also been observed several times (Burla and Giitz 1965; Krimbas and Alevizos 1973; Burla et al. 1986). Habitat choice of the various karyotypes of this species has been described by Kekic et al. (1980), Shorrocks and Nigro (198 I), and Cabrera et al. (1985). There also exist differences in the daily activity of the various allozyme variants (Cabrera et al. 1985). Furthermore, relationships between the different genetic traits mentioned above have also been discovered, as between gene arrangements and quantitative traits (Prevosti 1967) and, mainly, as nonrandom associations between allozyme variants and chromosomal inversions (Zouros et al. 1974; Loukas and Krimbas 1975; Loukas et al. 1979; Charlesworth et al. 1979; Pinsker and Sperlich 1982; Cabrera et al. 1983; Prevosti et al. 1983). The strength of this gametic disequilibrium may even show seasonal changes in some populations (Fontdevila et al. 1983). In spite of this large amount of information, the relative importance of historical or selective factors for the determination of the macro- and microstructure of the D. subobscura populations remains unclear. Mitochondrial DNA (mtDNA) variation has recently become a favorite tool for the study of intraspecific population structure on a fine scale, because of its considerable polymorphism and its strong sensitivity to founder events or population subdivision. The main reasons for this are the maternal mode of inheritance and the lack of recombination (for a review, see Wilson et al. 1985). Hence, it was of interest to study the mtDNA polymorphism of those continental and insular populations of D. subobscura that had already been analyzed previously, with respect to chromosomal and allozyme variability, by other authors and ourselves (Lakovaara and Saura 197 1; Saura et al. 1973; Prevosti 1974; Pinsker et al. 1978; Pinsker and Sperlich 1979; Cabrera et al. 1980; Krimbas and Loukas 1980; Sperlich et al. 1980; Larruga et al. 1983; Prevosti et al. 1983; Cabrera et al. 1985). Our main aim was to find out whether the pattern of mtDNA variability fits that of inversion polymorphism and those characters that are determined by nuclear genes. Material and Methods
A total of 26 1 isofemale lines of Drosophila subobscura from the Old World were analyzed. Some populations were sampled twice in order to detect temporal variation, and, in one locality (Raices), two ecologically distinct areas were examined at the same time to obtain information on microgeographic differentiation. The localities sampled (fig. 1) range from the northern border of the distribution area of the species in Europe (Sweden and Scotland) to the southern margin (the Canary Islands and Morocco). The regions and populations sampled were as follows: Tenerife, Canary Islands-l 5 lines from Calderetas, collected in 1985, and 49 lines from Raices: 36 sampled in 1983 (R830 = 18 lines from an open site and R83g = 18 lines from a shaded gully) and 13 lines (R87) sampled in 1987; Madeira-47 lines from Ribeiro and 20 lines from Poiso, sampled in 1987; Morocco- 14 lines from Chechouan, collected in 1986; Azores-two lines from the Island of St. Maria, sampled in 1979; Iberian Peninsula-19 lines from Villares and 16 lines from Escorial, sampled in 1985; France: three lines from Paris, collected in 1987; Switzerland-l 3 lines from Zurich, collected in 1987 (in addition, three more lines were analyzed from the same locality but only for the restriction enzyme HaeIII; Germany-37 lines from Tiibingen (13 lines sampled in 1986 and 24 lines sampled in 1987); Italy-one line each from
126 Afonso et al.
Formia, Ponza, and Sicily (all three collected in 1977); Sweden-six lines from Gavle, three lines from Uppsala, and three lines from Sundsvall (all six collected in 1987); USSR-one line from Caucasus, collected in 1979. mtDNA isolation from adult Drosophila flies, enzymatic digestion by restriction enzymes, electrophoresis, and the registration of fragment-length pattern is as described by Afonso et al. (1988). Eleven restriction enzymes were used in the present study: Two of them (HaeIII and HpaII) recognize 4-bp sequences. The other nine (BarnHI, EcoRI, EcoRV, HindIII, PstI, PvuII, SacI, XbaI, and XhoI) recognize 6-bp sequences. To facilitate comparisons, the different restriction patterns of a given enzyme will be designated by capital letters (figs. 2 and 3), in accordance with the protocol of Latorre et al. ( 1986), but the nucleomorphs deduced from the patterns of all enzymes together (Nei and Tajima 198 1) will be denoted by arabic and not by roman numbers (figs. 35), because of the large number of different nucleomorphs found in the present survey. The construction of the restriction map of the mtDNA molecule was carried out in three different ways: double digestions with pairs of enzymes that give only a few restriction fragments, partial digestions for enzymes with numerous and/or small fragments, and filter hybridization of totally digested DNA with the already mapped clones of D. melanogaster 62F9, H, B, and M.3 [constructed by Garesse (1988)] and of D. yakuba pDYHB, pDYHC, and pDYHD, [prepared by Clary and Wolstenholme (1985)], by using photobiotinated probes (Forster et al. 1985). Restriction-fragment sizes were determined by using both lambda DNA cut with Hind111 and pBR322 DNA cut with HaeIII as reference markers and by applying the Schaffer and Sederoff ( 198 1) program for length estimation. The intra- and interpopulation nucleotide diversity was estimated with equations (18) and (3 1) of Nei and Tajima (198 1). The partition of the total genetic variation in subdivided subpopulations was obtained by the method of Nei (1982), and the calculations of the amount of genetic differentiation between nucleomorphs were done by using equations (1 l), (19), and (25) of Nei and Tajima ( 1983). Phylogenetic trees were constructed using the unweighted-pair-group method (Sneath and Sokal 1973) or the neighbor-joining method (Saitou and Nei 1987). Results
Of the 11 enzymes used in the present survey, only one (XhoI) did not cut the mtDNA molecule in any of the strains, three (BarnHI, PvuII, and XbaI) produced the same pattern in all strains, and the other seven were “polymorphic” at least in one of the strains (fig. 2). The largest number of restriction fragments that a simple digestion yielded was five for the restriction variants EcoRI C, HpaII C, and HpaII D. In the map construction (fig. 2), the smallest fragment detected by our method was 300 bp in length when Sac1 was used, but in the ordinary screening program of the population studies the smallest fragments which could be reliably discovered were in the range of 500 bp. The mean length of Drosophila subobscura mtDNA was estimated to be 15,900 + 300 bp. Since the estimation of the length of long fragments in the gel is rather imprecise, an indirect approach for the estimation was used. There are three conservative Hind111 sites in Drosophila mtDNA that are also present in D. subobscura (figs. 2 and 6). The two large fragments of Hind111 digestion are 8,450 and 5,300 bp long (Clary and Wolstenholme 1985; Solignac et al. 1986b). The remaining fragment, containing the AT-rich region, is at most 2,150 bp long in D. subobscura. The total length must therefore be 15,900 bp. This is considerably shorter than the mtDNA of D.
mtDNA Variation in D.subobscura 127
Barn Eco Eco Hae Hind Hpa Pvu Pst sac Xba
HI RV RI III III II II I I I
Eco
RV
I#] 'YVVV 1.nnn
J
II
IVV
L cnn UJVV I1
Lnn
I* uvv
c
RI
Hae III
10550
c inn
I
II_
11550
0000
Eco
’ 3501
194
11oq
14OOl ]
14750 7100
1804
4300
4600
100
and
II aq
115ol
4350
ll4OoI
9900
I
1900 I
B C 0 C 0
0 200
15900
A A A A A A A A A A
F 0 E
9900
Hind III
B C 0 E
Hpa II
B C D
Pst
I
Sac I
15900
I#1
4000
B
I I 7300
I
2 800
I
6500
B
FIG. 2.-mtDNA cleavage maps from the 10 enzymes that cut the molecule at least once. The 17 restrictionenzymes producing the fragments indicated are listed on the left side; the capital letters designating the respective restriction morphs are listed on the right side.
melanogaster or D. yakuba. Yet, in regard to the AT-rich region, size variations between species of the same genus are well known for Drosophila (Hale and Beckenbach 1985;
Hale and Singh 1986; Solignac et al. 1986a). Thirty-eight restriction sites were detected. Half of them proved to be unvaried; the other half were polymorphic, as can be seen from figure 6, where our D. subobscura
D
x
2
AABAABB
1000 Can
10
AABABBB
Can
4
CABAABB
Can
12
AABBABB
Can
1
AABAAAB
Can Can
6
AAAAAAB
7
AABAAAA
Can
9
AABAABA
Can
5
CABAACB
Can
11
AABAACB
Can
3
AAAAACA
Cos
8
AABAACA
Mmc
AACAACA
Hor
15 AAAAADA
23
Mad
21
AAAAACD
Ger
16
AAABACA
Mad
18
ADAAAAA
Spa
20
AAAAAAC
Ger
13
AAAAAEA
Ger
22
ABAAAAA
Ger
14
ACAAAAA
Mad
17
AADAAAA
Mad
13
AAAAAAA
Cos
24
DAAAAAA
Ger
2 AABAABB 12 AABBABB IO AABABBB 4 CABAABB
2.48
0.37 I
0.64
4r-
9 AABAABA
.69 I
5 CABAACB 0.93 0.36
11 AABAACB 1 AABAAAB 7 AABAAAA 8 AABAACA 3 AAAAACA 16
AAABACA
15
Can Can Can Can Can Can Can Can Can Mmc cos Mad
AAAAADA 21 AAAAACD 23 AACAACA 6 AAAAAAB
Mad Ger
17
Mor Can
AADAAAA
Mad
14
ACAAAAA 18 ADAAAAA
Mad
22 ABAAAAA
Ger
20 AAAAAAC 19 AAAAAEA 24 DAAAAAA
Ger
13
cos
AAAAAAA
Spa
Ger Ger
FIG. 3.-Phylogenetic trees of the 24 nucleomorphs obtained from the 6 values given in fig. 4. The various nucleomorph designations (AABAAB, etc.) are the same as in table 2 and correspond to the different restriction patterns of the polymorphic restriction enzymes in the sequence of table 1. Shown are the UPGMA method (a) and the neighbor-joining method (b) (the figures over each branch are the patristic distances multiplied by 1,000). Can = Canary Islands; Mad = Madeira; Mor = Morocco; Ger = Germany; Spa = Iberian Peninsula; Cos = cosmopolitan; Mmc = Madeira, Morocco, and Canary Islands.
mtDNA Variation in D. subobscura
129
restriction map is compared with the genetic map of D. yakuba by Clary and Wolstenholme ( 1985). Only one of the 26 1 isofemale lines tested proved to be heteroplasmic. This line comes from the 1986 sample from Tiibingen and contains two types of mtDNA, differing in total mtDNA size. One mtDNA type corresponds to that found in all the other lines, whereas the other is considerably shorter (N 11 kb). The frequencies of different restriction patterns for the seven polymorphic enzymes for all the populations studied are shown in table 1. The enzymes EcoRI, EcoRV, PstI, and Sac1 show a cardinal pattern that is the same in all the geographic samples with rare local variants. HpaII is consistently polymorphic only in the Canary Islands populations where variant B is the most frequent. In all other regions variant A is the only or the most important pattern. A similar situation exists for the enzyme HindIII; again the most frequent restriction pattern (B), found in the Canary Islands, is not the one that is fixed on the European continent. In this case, however, the nearest geographic populations to the Canary Islands, i.e., Morocco and Madeira, have also the Canary Islands variant, although at low frequencies. The only enzyme that is polymorphic in all populations studied is HaeIII. Once more, the most common variant of the Canary Islands (B) is absent in all other populations, whereas the types common elsewhere (A and C) are present only at relatively low frequencies in the Canary Islands. Furthermore, the frequency distribution of these common variants displays a northsouth clinal variation, ranging from frequency values as low as 0.083 for variant A in Sweden to 0.526 in one of the Iberian Peninsula populations (Villares). In spite of the temporal variation existing in the two samples from Tubingen, the correlation (when a moment-product r is used) between the frequency of pattern A and latitude is statistically significant when all samples are used (r = -0.874; df = 9; P < 0.001) and is still so, though at a lower level, when the three Swedish samples are pooled and when the small French sample is excluded (r = -0.777; df = 5; P < 0.05). The percentages of the 24 different nucleomorphs detected in the various geographic lines are given in table 2. The overall distribution pattern is the same as that deduced from restriction fragments of single enzymes. There are two cosmopolitan nucleomorphs (3 and 13) with high frequencies in all the populations, except in the Canary Islands. On the other hand, the most common nucleomorph (2) in the Canary Islands is absent from all the other populations. Nucleomorph 8 is present only in North Africa, the Canary Islands, and the Madeira Islands. The rest of the nucleomorphs are unique to single populations. Again, the difference between the Canary Islands populations and the others is evident. The nucleotide diversity, rr, has been proposed as a fundamental measure of the amount of DNA variability in a given population (Nei and Li 1979). It is defined as the average per-site number of nucleotide differences between two randomly chosen DNA sequences. The existence of 4 bp- and 6 bp-specific enzymes has been taken into account. The value from our data for the total sample including all populations is 0.008 1. This can be taken as an estimate of the mtDNA diversity of D. subobscura in its original distribution area. The values, multiplied by 102, for the different regions from which enough single female lines were studied are presented in the diagonal of table 3. The most extreme values are those from Tenerife (0.47) and Sweden (0.07). In the same table, the estimates of DNA divergence (6), multiplied by 102, between regions are presented above the diagonal. This 6 value relates to the average number of nucleotide substitutions for a randomly chosen pair of mtDNA molecules from two different populations minus the arithmetic mean of the average numbers of nu-
l-l r-l
m
I
I l-4
d
+I
mcoua3w~wwbcQ~~w~wcn
4
i-l +I +I i-l 4-I +I ~muwca-G3m-mNNa3-cz~
w~~~w~uw~~w~ww~w~b
-I-II+ A
i-l
+I
+I I-Id-l-l
+I
+I
i-l
-f-l +I +I
4
+I
4
+I
+I
+I
+I I-f
+I
+I I-Id
+I 4
i-l
l-4
+I -:-I +I 4-l i-l +I -+I +I 4-I -l-l
I+-
i-l +I i-l 91 +I +I +I +I +I Qc4PJ-a3cJuanJcJco-Qc3Ncol-lm
+I +I
~~wwu~~wouwuuwr-wwuwu~ I +I +I +I +I +I 91 4-I +I g ~~~~u~~~~u~uu~=~~u~u~
8
co
-
m H
mtDNA Variation in D. subobscuru
131
FIG. 5.-Network of the 24 mtDNA nucleomorphs detected. The diagnostic site(s) that have mutated between each of the linked pairs of nucleomorphs are depicted; b = BumHI; i = EcuRI; e = EcoRV; a = H&II; h = HindIII; p = HpaII; t = PstI; v = PvuII; s = SacI; and x = XbuI. Each arrow indicates the gain of a restriction site.
cleotide substitutions that are found for pairs of mtDNA molecules from the same populations because of the effect of polymorphism. For the calculation of the 6 values, the frequencies of each nucleomorph in each region are taken into account (Nei and Li 1979; Nei and Tajima 198 1). In the present case, the nucleotide divergence was estimated, according to the method of Becker et al. (1988), for 4 bp- and 6 bp-specific restriction enzymes separately and was averaged by weighting each n or 6 value by the number of nucleotide sites detected by each group of restriction enzymes. As can be seen from table 3, the Canary Islands prove to be the only region that is well differentiated from the rest in the distribution area of D. subobscura, with an average divergence of 6 = ( 1.19 + 0.04)( 10e2). In contrast, the divergences between the other regions are practically nonexistent, though the geographic distances and the natural barriers are substantial for some pairs of populations. It is also possible to estimate values of divergence between the different mtDNA nucleomorphs. Figure 4 presents the values (X 103), between pairs of mtDNA nu-
xh
cl
is
P
si
I
a
s
I
I
e
e t
I
I’
ha
I
I
h
1
‘ep a
I
t
pv
I
I
I
a
i
s
e
p
sb
h
x
I
1
1
Ph
e
h
I
’ I
P
,
. A
E
RNA
LrRNA
L
“6” 2: iz
0
5
9
I
I
ND1
I
I
Cytb
I
ND4
co11
ND 5
co I
ND 2
A+T 1
10 I
I
I
I
16 I
I
I
I
I
1
kb FIG. 6.-Distribution of the restriction sites observed in the functional regions of the mtDNA molecule of Drosophila subobscura (above) aligned to the genetic map of D.yakuba(middle) of Clary and Wolstenholme (1985). Each vertical line corresponds to a restriction site, and the lengths of the lines are proportional to the number of those mtDNA types that were cut at this specific site. The lengths of the various restriction fragments can be deduced from the kilobase (kb) scale given below. b = BumHI; i = EcoRI; e = EcoRV; a = HueIII; h = HindIII; p = HpuII; t = MI; v = PvuII; s = SucI; and x = XbuI.
Table 1 Restriction-Pattern
Frequencies of Polymorphic Enzymes in the Samples Studied
Germany
Tenerife Madeira RESTRICTION PATTERN EcoRI: A ............ C ............ D ............ EcoRV: A ............ B ...............
c .................. D ............ t; w
HueIII: A ............ B ............ c ............ D ............... E ............... HindIII: A ............ B ............ c .................. D ............ HpuII: A ............ B ............ c ............ D ............ PstI: A ............ B .................. SacI: A. ........... B ............... Total no. in sample .....
R830
0.83 0.17 ... 1.00
R83g
1.00
R87
1.00
1.00
...
... .. .
...... ...
Calderetas
1.00
1.00
... ... ...
... ... ...
b:d2
0.07
0.30
0.15 0.77 0.08
G3 0.07
...... ...
...
... ...
;:&
0.08 0.92
0.07 0.93
... ...
... ...
0.17 0.83
0.31 0.69
0.07 0.93
...
...
...
... ...
1.00
1.00
0.92 0.08
......
.........
18
0.98
1.00
0.06 0.83 0.11
1.00
... ...
... ...
0.17 0.67 0.17
0.94 0.06 18
Italy
Villares
Escorial
France
1.00 ... ... 0.94 ...
1.00
... ... ...
b:d6 0.25
0.67
0.53
0.31
b:93 ... ...
b:88 0.02
0.79 0.14 0.07
0.94 0.04
...
b:d2
b:;5
b:j3
b:47
b:89
...
...
...
... ...
...
0.44
0.33
Tiibingin 1986
Tiibingen 1987
Scotland
Sweden
1.00 ... ...
0.96
1.00
1.00
b:d4
... ...
... ...
0.92 0.08
1.00
1.00
... ...
1.00 ... ... ...
... ... ...
... ... ...
0.46
0.29
0.20
0.08
b:46
b:;l
Go
G2
b:d8
,.. ...
... ...
*. . ...
1.00
1.00
1.00
1.00
... ... ...
... ... ...
... ... ...
0.85
1.00
1.00
1.00
b:d8 0.08
... ... ...
... ... ...
... ... ...
1.00
1.00
... ... ...
1.00
1.00
...
...
0.98 0.02 1.00
1.00
1.00
...
... 14
47
... ...
... ... ...
... ... ...
1.00
0.67
1.00
1.00
1.00
b:;6 ...
... ... . ..
... ... ...
15
Switzerland
...
... 13
Poiso
1.00
J.00 ... ...
1.00
......
0.17 0.83
Ribeiro
... ...
0.11 0.89
Morocco
Iberian Peninsula
1.00
...
1.00
1.00
1.00
1.00
...
...
...
...
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
...
...
...
...
...
...
...
...
...
20
3
19
16
13
13
24
10
12
OVERALL
0.99 0.01 0.00 0.99 0.00 0.00 0.00 0.25 0.20 0.55 0.00 0.00 0.75 0.24 0.00 0.00 0.79 0.20 0.00 0.00 0.99 0.01 1.00 0.00 261
.......
*
*
.
*
*
*
*
:*:::::::::::
.t-?.
..........
*
........
‘Q\
.I-.
*
.
. .,'
...........
........... ...........
:;u”
.q
1%
.q :
:
:
........
:
:
:
:
:
.q
:
:
:g
:
:
.
:
:
:
........ : :
.m.
.cq
:
:
:
:
:
.
:
:
*
.
:-
:g
.........
.
. :
......... .........
:
. : ‘:
.cq :g “_
. : ‘7
. : “.
. :
:
It-l-t-r-
. :
.
. :
. :
. :
. :
. :
:
:
:
. :
.
:
:
.w_ 1%
.
.
1.
:
........... ........... ...........
:
.q :g
: : : : : : : : :g
........ :
.
:
. . . . . . . . . . .
:g
.
.
.
: : :
.
.o..
.
:z
.ty . :w::::::
.
.
:
.
:
.
.
.
:
.
.
.
.
.
:
.
.
.
.
.
:
.
.
.
.
.
:
.
.
.
.
.
.
.Iy
.q
.
.
. .
:
.
:
.
N _
2
24
m
m
m _
w
_ Q\
VI
:
.
d
s
z :
.
:
. . . . . . . . . . .. .. .. .. .. .. .. .. .. .. ..
:
.
.-.
.
:
.
:
.
.q
:
:
.
.
.
.
.
.
:
.
*v,mritvc\1
:
.\9-*-yy ..
:
.
‘CI
.
:
:
.
:
.
:
:
.
16 :
.
:
.*
. . . . . . . . .q
. .y . :s
.
. .c‘!
1%
.w*.
.
.
.
:
.
.
.
.
.
: : : : : : : : :m::::::::::: m
.
.
.
.
.w*
.
.
.
:s : : : : : : : : :g:::::::::::
.
.cq
.
.
:z
.
.
.
.
:
:
.
.
:-
.I-?.
:- _
.q..
.-.
: : : : : : : : :z .
. .
:c‘
.q
.
. .
. .. ..
.
. :
:z .‘:.
:z
.‘?.
.
. .
. .
.:;=...
.
Table 3 Nucleotide Diversity and DNA Divergence
t; VI
Tenerife ....... Madeira ....... Morocco ...... Iberian Peninsula Germany ...... Switzerland .... Scotland ....... Sweden ........ Overall ......
. .. . ..
I
Tenerife
Madeira
Morocco
Iberian Peninsula
Germany
Switzerland
Scotland
Sweden
0.47
1.15 0.25
1.20 0.02 0.21
1.02 0.01 0.06 0.24
1.28 0.01 0.06 0.00 0.42
1.14 0.00 0.03 0.01 0.00 0.20
1.22 0.00 0.02 0.02 0.02 0.01 0.15
1.32 0.02 0.01 0.05 0.06 0.02 0.01 0.07
NOTE.-Data are ‘Kvalues (on the diagonal) and 6 values (above the diagonal), multiplied by 102, in and between the sampled regions (for further explanation of * and 6, see the text).
Overall
0.81
136
Afonso et al.
cleomorphs, calculated by the maximum-likelihood method of Nei and Tajima ( 1983). The maximum distance value (25.6) of figure 4 is that between nucleomorph 12, endemic to the Canary Islands, and nucleomorph 24 from Germany. Within regions, the maximum distance (20.5) is that between nucleomorphs 3 and 12 of the Canary Islands, compared with the distance value ( 16.7) between nucleomorphs 2 1 and 24, which is the maximum distance between the nucleomorphs for all European populations. Figure 3 shows the estimated genealogical relationships among the various nucleomorphs. Both phylogenetic trees generally reflect the same pattern of relationship. There are two main clusters, one of them formed by nearly all the endemic nucleomorphs from the Canary Islands and the other composed of all the cosmopolitan and non-Canary Islands nucleomorphs. In figure 5 the 24 nucleomorphs are arranged in such a way that the connecting lines minimize the total number of restriction-site changes required to move from one nucleomorph to the other. All nucleomorphs directly connected differ from each other only by a single mutational step, with exception of the pairs 13- 18 and 13-24, which are separated by two nucleotide changes. Yet, some of the relationships in the minimal network of nucleomorphs involve successively the mutational appearance and disappearance of the same restriction sites of the enzymes HaeIII, HindIII, or HpaII (fig. 5). This fact somewhat confuses the genealogical network between nucleomorphs; e.g., the most common nucleomorph (2) of the Canary Islands can be connected with the central nucleomorph 13 in three ways, via nucleomorphs l-6,9-7, and l-7 always gaining and losing identical restriction sites. The phenomenon could be explained by two hypotheses: parallelism (the same mutation occurring more than once) or recombination. The latter hypothesis has a very low probability because of the paucity of recombinational events in animal mtDNA (Wilson et al. 1985). Another way to analyze the structure of D. subobscura populations is by partitioning of the total mtDNA variation in successive subdivisions. When Nei’s (1982) method for analysis of gene diversity is applied to our data, the values of table 4 are obtained. Of the total mtDNA diversity, 61.23% appears to arise from differences among regions, 38.40% from nucleotide diversity within samples, and only 0.25% from differentiation within regions. Expressed as part of the total diversity, the value for the temporal differentiation effect is only 0.12%. However, when the temporal subdivision is considered in only the localities where data for different years exist, the differentiation in Ttibingen (10.69%) is conspecious compared with its absence in Raices (Canary Islands). Discussion
Our estimation of the size of the mtDNA of Drosophila subobscura is 15.9 kb. This is less than the values calculated by Nigro and Powell (1985) and Latorre et al. Table 4 Partitioning of mtDNA Variability
0.008 1
0.3840
0.0012
0.0025
0.6123
NOTE.-Data are expressions of variability of Drosophih subob.scuru according to temporal and spatial components of Nei (1985). 1~,= Total diversity; ySrl = individual variation; ysr2 = temporal differentiation; ySr3= differentiation among localities; and yS,4 = differentiation among regions.
mtDNA Variation in D. subobscura
137
(1986). The reason for the discrepancy is most probably the uncertainty in the estimation, on the basis of the position of the bands in the gel, of long fragments. Smaller fragments can be measured much better. We are quite sure that the smallest fragment of Hind111 is 12,150 bp and that the other two fragments are 8,450 bp and 5,300 bp. The latter figures are consistent with the values of Clary and Wolstenholme (1985), who have sequenced the mtDNA of D. yakuba. The average nucleotide diversity, ‘Tc,was 0.008 for D. subobscura in our study. This is close to the value found by Latorre et al. (1986), using only data from Old World strains of D. subobscura (2 = 0.0 11). The majority of estimates for other species of Drosophila (Shah and Langley 1979; Hale and Beckenbach 1985; Hale and Singh 1987) fall within the range 0.002-0.0 11. Similar values were found by us for our different populations (table 3). The most significant result of the present study is the large nucleotide divergence between the Canary Islands population and the other populations (table 3 and figs. 3 and 4). The average 6 (0.0 119 + 0.0004) is at the same level as that between the sibling species D. pseudoobscura and D. persimilis. Following Becker et al. ( 1988) we have calculated a 6 value of 0.0 12 on the basis of Hale and Beckenbach’s (1985) data. Yet, there is not the slightest indication for the existence of any sexual isolation between Canary Islands and European strains of D. subobscura (Constanti et al. 1986). Similar significant differentiation of the Canary Islands populations has been found with respect to chromosomal polymorphism and allozyme variability (Prevosti et al. 1975; Larruga et al. 1983). The remarkable genetic divergence of the populations from Canary Islands can be explained by geographic isolation that reduces the gene flow from the continent nearly to zero. Yet, the question remains whether the colonization of Canary Islands by D. subobscura is a recent or an ancient event. Prevosti (1974) concluded, from the similarity of chromosomal inversion polymorphism in the Canary Islands and Madeira and from the striking difference between them and the population in Morocco, that the polymorphism in the islands is an old relict polymorphism that has reached a less evolved stage on the continent. Data on allozyme polymorphism, on the other hand, demonstrate that all the different Canary Islands populations are genetically similar (Cabrera et al. 1980) but the Madeira populations are more related to those of the continent than to Canary Islands populations (Larruga et al. 1983). The data from mtDNA divergence show the same picture (table 3). Yet, there are other reasons for believing that Canary Islands populations are indeed very ancient. The maximum distance between nucleomorphs in the Canary Islands populations is 0.02 1, nearly the same magnitude as that found between all nucleomorphs in the present study (0.026). We can convert this molecular divergence value into time, assuming that t = (l/k) X 6 X 0.5, where k is the substitution rate per site per year. For Drosophila mtDNA the k estimates are between 1 X lo-’ and 30 X lo-’ (Solignac et al. 19863). This means that the two most divergent nucleomorphs have diverged from a common ancestral molecule between 10.2 Myr ago (Mya) and 0.3 Mya. Another result that seems to confirm the old age of the colonization of the Canary Islands by D. subobscura is the large number of different endemic nucleomorphs found in the insular populations compared with other insular or continental populations of similar sample sizes (table 2). If all these assumptions are true, it is possible that the populations of the Canary Islands represent the ancestral composition of the originally African populations of D. subobscura, whereas actual continental populations might have originated more recently from a few founders after the last glacial period. The existence, in the Canary
138 Afonso et al.
Island archipelagos and in Madeira, of a relict Tertiary flora which disappeared from the continents supports this hypothesis. Later, the gene flow from Europe might have affected the composition of the populations of Madeira but not those of the Canary Islands. Though a relatively high homogeneity for nucleomorph frequencies exists for the European populations, it is worth mentioning that northern, marginal populations appear less polymorphic than central ones (tables 1 and 2). Precisely the same pattern of geographic differentiation has been found for chromosomal inversion polymorphism. Heterozygosities for allozyme variability, on the other hand, are similar in all populations. The different patterns can be explained, if we assume that marginal populations are repeatedly reduced to a small number of individuals. Because of its maternal inheritance, mtDNA is a stronger detector of bottlenecks than are nuclear genes (Lansman et al. 198 1; Wilson et al. 1985). For chromosomal gene arrangements, however, strong natural selection must be considered. Another unexpected result is the existence of a north-south clinal variation in the frequencies of mtDNA types A and C of the polymorphic enzyme Hue111 (table 1). A similar geographic pattern is found for some quantitative traits (Prevosti 1955; Pfriem 1983), chromosomal gene arrangements (Krimbas and Loukas 1980), and certain allozyme variants (Pinsker and Sperlich 1979; Prevosti et al. 1983). Both the hypothesis that the processes are historical and the assumption that adaptive mechanisms have determined the geographic structure of European populations have been proposed (Krimbas and Loukas 1980). Recently, the existence of similar clinal patterns for both chromosomal polymorphism and wing length in the New World Drosophila have reinforced the adaptation hypothesis (Prevosti et al. 1985, 1988). The mtDNA cline found here would favor the historical explanation. Yet, no clinal variation of nucleomorph frequencies has ever been reported from other investigations. New studies on the geographic distribution of mtDNA polymorphism are needed to find out whether this mtDNA frequency cline is real and permanent. In spite of the high dispersal capacity of D. subobscura (Loukas and Krimbas 1979; Serra et al. 1987), in populations of this species there are some indications of microspatial and temporal differentiation (Kekic et al. 1980; Shorrocks and Nigro 198 1; De Frutos and Prevosti 1984; Cabrera et al. 1985) similar to those found in other species of Drosophila (Stalker 1976; Taylor and Powell 1977; Barker et al. 1986). Samples taken in different years at the same locality (tables 1 and 2) showed the existence of temporal fluctuations in nucleomorph frequencies in Tiibingen ( 10.69%) but a great stability in Raices/Tenerife (0%). This difference can be explained by the strong seasonal variations of ecological factors in Germany, which reduce the local population sizes more severely than does the mild and almost constant climate of the Canary Islands. Only 0.25% of the total diversity arises from differentiation between localities within regions. The relatively small sample sizes (table 1) could explain this result, but the distribution of nucleomorphs among samples makes us believe that a subdivision of populations is real. For example, the three samples from Raices have a similar size and contain five to six endemic nucleomorphs, whereas the sample from Calderetas, only 7 km away, has only one endemic nucleomorph. The same is true for the very close localities of Poiso and Ribeiro on Madeira. In conclusion, the results demonstrate that the wild populations of D. subobscura are structured with respect to mtDNA polymorphism at a geographical and local level in accordance with the many investigations with other genetic markers. Yet, the studies
mtDNA Variation in D. subobscura
139
of mtDNA variability complement the previous data significantly because they are gained on a molecular basis and because mtDNA transfer is governed by a different mode of inheritance. Acknowledgments
We gratefully acknowledge Dr. M. Nei for his valuable suggestions and for sending us the programs for calculating 6 values and for the neighbor-joining method. We are also indebted to Drs. Garesse and Wolstenholme for providing us with their mtDNA clones of Drosophila melanogaster and D. yakuba, respectively. We also thank C. Rehm for typing the manuscript and K. Stogerer for providing some of the drawings. Dr. A. Saura has collected D. subobscura for us in Sweden. This research was supported by CAICYT grant PR83-0396 to V.M.C. LITERATURE CITED AFONSO,J. M., J. PESTANO,and M. HERNANDEZ.1988. Rapid isolation of mitochondrial DNA from Drosophila adults. Biochem. Genet. 26:38 l-386. BARKER,J. S. F., P. D. EAST, and B. S. WEIR. 1986. Temporal and microgeographic variation in allozyme frequencies in a natural population of Drosophila buzzatii. Genetics 112:57761 1. BECKER,I. I., W. S. GRANT, R. KIRBY, and F. T. ROOB. 1988. Evolutionary divergence between sympatric species of southern African Hakes, Merluccius capensis and M. paradoxus. II. Restriction enzyme analysis of mitochondrial DNA. Heredity 61:2 l-30. BURLA, H., and G. GOTZ. 1965. Veranderlichkeit des chromosomalen Polymorphismus bei Drosophila subobscura. Genetica 36:83- 104. BURLA, H., H. JUNGEN, and G. B%HLI. 1986. Population structure of Drosophila subobscura: non-random microdispersion of inversion polymorphism on a mountain slope. Genetica 70:9-15. CABRERA,V. M., A. M. GONZALEZ,and A. GULL~N. 1980. Enzymatic polymorphism in Drosophila subobscura populations from the Canary Islands. Evolution 34:875-887. CABRERA,V. M., A. M. GONZALEZ,M. HERNANDEZ,J. M. LARRUGA,and M. MARTELL. 1985. Microgeographic and temporal genetic differentiation in natural populations of Drosophila subobscura. Genetics 110:247-256. CABRERA,V. M., A. M. GONZALEZ,J. M. LARRUGA,and C. VEGA. 1983. Linkage disequilibrium in chromosome A of Drosophila subobscura. Genetica 61:3-8. CHARLESWORTH,B., D. CHARLESWORTH,and M. LOUKAS. 1979. A study of linkage disequilibrium in British populations of Drosophila subobscura, with an appendix by K. Morgan. Genetics 92:983-994. CLARY, D. O., and D. R. WOLSTENHOLME.1985. The mitochondrial DNA molecule of Drosophila yakuba: nucleotide sequence, gene organization, and genetic code. J. Mol. Evol. 22: 252-271. CONSTANTI,P. M., G. RIB& and A. PREVOSTI. 1986. Sexual isolation between populations of Drosophila subobscura. I. European strains. Genet. Iber. 38:2 13-22 1. DE FRUTOS, R., and A. PREVOSTI. 1984. Temporal changes of chromosomal polymorphism in natural populations of Drosophila subobscura. Genetica 63: 18 I- 187. FONTDEVILA,A., C. ZAPATA, G. ALVAREZ,L. SANCHEZ,J. M~NDEZ, and I. ENRfQuEz. 1983. Genetic coadaptation in the chromosomal polymorphism of Drosophila subobscura. I. Seasonal changes of gametic disequilibrium in a natural population. Genetics 105:935-955. FORSTER, A. C., J. L. MCINNES, D. C. SIUNGLE, and R. H. SYMONS. 1985. Non-radioactive hybridization probes prepared by chemical labelling of DNA and RNA with a novel reagent, photobiotin. Nucleic Acids Res. 13:745-76 1. GARESSE,R. 1988. Drosophila melanogaster mitochondrial DNA: gene organization and evolutionary considerations. Genetics 118:649-663.
140
Afonso et al.
HALE, L. R., and A. T. BECKENBACH.1985. Mitochondrial DNA variation in Drosophila pseudoobscura and related species in Pacific Northwest populations. Can. J. Genet. Cytol. 27: 357-364.
HALE, L. R., and R. S. SINGH. 1986. Extensive size variation and heteroplasmy in mitochondrial DNA among geographic populations of Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 83:88 13-88 17. . 1987. Mitochondrial DNA variation and genetic structure in populations of Drosophila melanogaster. Mol. Biol. Evol. 4:622-637. KEKI~, V., C. E. TAYLOR,and M. ANDJELKOVI~.1980. Habitat choice and resource specialization by Drosophila subobscura. Genetika (Yugoslavia) 12:2 19-225. KRIMBAS, C. B. 1965. The genetics of Drosophila subobscura populations. I. Inversion polymorphism in populations of southern Greece. Evolution l&541-552. KRIMBAS,C. B., and V. ALEVIZOS. 1973. The genetics of Drosophila subobscura populations. IV. Further data on inversion polymorphism in Greece: evidence of microdifferentiation. Egypt. J. Genet. Cytol. 2: 12 l-l 32. KRIMBAS,C. B., and M. LOUKAS. 1980. The inversion polymorphism of Drosophila subobscura. Pp. 163-234 in M. Hecht, W. C. Steere, and B. Wallace, eds. Evolutionary biology. Vol. 12. Plenum, New York. LAKOVAARA,S., and A. SAURA. 197 1. Genie variation in marginal populations of Drosophila subobscura. Hereditas 69~77-82. LANSMAN,R. A., R. 0. SHADE, J. F. SHAPIRA,and J. C. AVISE. 198 1. The use of restriction endonucleases to measure mitochondrial DNA sequence relatedness in natural populations. III. Techniques and potential applications. J. Mol. Evol. 17:214-226. LARRUGA, J. M., V. M. CABRERA,A. M. GONZALEZ, and A. GULL~N. 1983. Molecular and chromosomal polymorphism in continental and insular populations from the southwestern range of Drosophila subobscura. Genetica 60: 19 l-205. LATORRE,A., A. MOYA, and F. J. AYALA. 1986. Evolution of mitochondrial DNA in Drosophila subobscura. Proc. Natl. Acad. Sci. USA 83:8649-8653. LOUKAS, M., and C. B. KRIMBAS. 1975. The genetics of Drosophila subobscura populations. V. A study of linkage disequilibrium in natural populations between genes and inversions of the E chromosome. Genetics 80:33 l-347. . 1979. The genetics of Drosophila subobscura populations. X. A study of dispersal. Genetica 50: 127- 134. LOUKAS, M., C. B. KRIMBAS,and Y. VERGINI. 1979. The genetics of Drosophila subobscura populations. IX. Studies on linkage disequilibrium in four natural populations. Genetics 93: \ 497-523. NEI, M. 1982. Evolution of human races at the gene level. Pp. 167-18 1 in Human genetics. Part A: The unfolding genome. Alan R. Liss, New York. NEI, M., and W. H. LI. 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76:5269-5273. NEI, M., and F. TAJIMA. 198 1. DNA polymorphism detectable by restriction endonucleases. Genetics 97: 145- 163. . 1983. Maximum likelihood estimation of the number of nucleotide substitutions from restriction sites data. Genetics 105:207-2 17. NIGRO, L., and J. R. POWELL. 1985. Preliminary results on polymorphism in mitochondrial DNA of Drosophila subobscura. IXth European Drosophila Research Conference, Hungary (abstracts). PFXIEM,P. 1983. Latitudinal variation in wing size in Drosophila subobscura and its dependence on polygenes of chromosome 0. Genetica 61:22 l-232. PFRIEM, P., and D. SPERLICH. 1982. Wild O-chromosomes of Drosophila subobscura from different geographic regions have different effects on viability. Genetica 60:47-59.
mtDNA Variation in D.subobscura 141 PINSKER, W., P. LANKINEN,and D. SPERLICH. 1978. Allozyme and inversion polymorphism in a central European population of Drosophila subobscura. Genetica 48:207-2 14. PINSKER,W., and D. SPERLICH.1979. Allozyme variation in natural populations of Drosophila subobscura along a north-south gradient. Genetica 50: 16 1- 170. in Drosophila subobscura. II. Non-random associations . 1982. MDH-polymorphism between alleles and chromosomal inversions in natural populations. Z. Zool. Syst. Evol. Forsch 20:161-170. PREVOSTI,A. 1955. Geographical variability in quantitative traits in populations of Drosophila subobscura. Cold Spring Harbor Symp. Quant. Biol. 20:294-298. . 1964. Chromosomal polymorphism in Drosophila subobscura populations from Barcelona (Spain). Genet. Res. 5:27-38. . 1967. Inversion heterozygosity and selection for wing length in Drosophila subobscura. Genet. Res. 10:8 l-93. . 1974. Chromosomal inversion polymorphism in the southwestern range of DrosophiZa subobscura distribution area. Genetica 45: 1 1 l- 124. PREVOSTI,A., M. P. GARcfA, L. SERRA, M. AGUAD~~,G. RIBS, and E. SAGARRA. 1983. Association between allelic isozyme alleles and chromosomal arrangements in European pop ulations and Chilean colonizers of Drosophila subobscura. Pp. 17 1- 19 1 in M. C. RATAZZI, J. G. SCANDALIOS,and G. S. WHITT, eds. Isozymes: Current topics in biological and medical research. Vol. 10: Genetics and evolution. Alan R. Liss, New York. PREVOSTI,A., J. OCARA, and G. ALONSO. 1975. Distances between populations of Drosophila subobscura based on chromosome arrangements frequencies. Theor. Appl. Genet. 45:23 l241. PREVOSTI,A., G. RIBS, L. SERRA, M. AGUAD~, J. BALARA,M. MONCL~S, and F. MESTRES. 1988. Colonization of America by Drosophila subobscura: experiment in natural populations that supports the adaptive role of chromosomal-inversion polymorphism. Proc. Natl. Acad. Sci. USA 85:5597-5600. PREVOSTI,A., L. SERRA,G. RIB& M. AGUADB,E. SAGARRA,M. MONCL~S,and M. P. GARcfA. 1985. The colonization of Drosophila subobscura in Chile. II. Clines in the chromosomal arrangements. Evolution 39:838-844. SAITOU,N., and M. NEI. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425. SAURA, A., S. LAKOVAARA,J. LOKKI, and P. LANKINEN. 1973. Genie variation in central and marginal populations of Drosophila subobscura. Hereditas 75~33-46. SCHAFFER,H. E., and R. R. SEDEROFF. 198 1. Improved estimation of DNA fragment lengths from agarose gels. Anal. Biochem. 115: 113- 122. SERRA,L., G. PEGUEROLES,and F. MESTRES.1987. Capacity of dispersal of a colonizing species: Drosophila subobscura. Genetica 73~223-237. SHAH, D. M., and C. H. LANGLEY. 1979. Inter- and intraspecific variation in restriction maps of Drosophila mitochondrial DNAs. Nature 281:696-699. SHORROCKS,B., and L. NIGRO. 198 1. Microdistribution and habitat selection in Drosophila subobscura Collin. Biol. J. Linnean Sot. 16:293-301. SNEATH, P., and R. SOKAL. 1973. Numerical taxonomy. W. H. Freeman, San Francisco. SOLIGNAC,M., M. MONNEROT,and J. C. MOUNOLOU. 1986a. Concerted evolution of sequence repeats in Drosophila mitochondrial DNA. J. Mol. Evol. 24:53-60. . 1986b. Mitochondrial DNA evolution in the melanogaster species subgroup of Drosophila J. Mol. Evol. 23:3 l-40. SPERLICH,D., H. FEUERBACH-MRAVLAG,P. LANGE, A. MICHAELIDIS,and A. PENTZOS-DAPONTE. 1977. Genetic load and viability distribution in central and marginal populations of Drosophila subobscura. Genetics 86~835-848. SPERLICH,D., W. PINSKER, and P. PFRIEM. 1980. Inversion, allozyme and lethal frequencies in natural populations of Drosophila subobscura. Genetika (Yugoslavia) 12:9 1- 10 1.
142 Afonso et al.
STALKER,H. D. 1976. Chromosome studies in wild populations of D. melanogaster. Genetics 83:887-894.
TAYLOR, C. E., and J. R. POWELL. 1977. Microgeographic differentation of chromosomal and enzyme polymorphism in Drosophila persimilis. Genetics 85:68 l-695. WILSON, A. C., R. L. CANN, S. M. CARR, M. GEORGE, U. B. GYLLENSTEN,K. M. HELMBYCHOWSKI,R. G. HIGUCHI, S. R. PALUMBI, E. M. PRAGER, R. D. SAGE, and M. STONENIUNG. 1985. Mitochondrial DNA and two perspectives on evolutionary genetics. Biol. J. Linnean Sot. 26:375-400. ZOUROS, E., C. B. KRIMBAS, S. TSAKAS, and M. L~UKAS. 1974. Genie versus chromosomal variation in natural populations of Drosophila subobscura. Genetics 78: 1223- 1244. MASATOSHI NEI, reviewing editor
Received March 17, 1989; revision received September 20, 1989 Accepted September 22, 1989
