inverted and standard arrangements in inversion
heterozygotes may explain why the introgressed
region is coincident with the entire 22-Mb 3La
The introgression profile was consistent between
samples from natural populations and the reference genome assemblies, involving the same species pairs and the same chromosomal locations
(supplementary text S4 and fig. S24). Moreover,
based on population samples from multiple geographic locations in Africa, patterns of introgression between An. arabiensis and An. gambiae +
An. coluzzii, and between An. merus and An.
quadriannulatus, are similar across their geographic range (fig. S25 and table S10). These findings refute the possibility that introgression was
an unnatural artifact of colonization and laboratory maintenance of multiple species. The lack of
geographic variation in patterns of introgression
also suggests that autosomal introgression occurred sufficiently long ago to have spread across
subpopulations. By contrast, mitochondrial DNA
(mtDNA) revealed patterns consistent with ongoing gene flow between An. arabiensis and An.
gambiae or An. coluzzii (supplementary text 3. 3
and fig. S22).
Transspecific inversion polymorphism
The unusually high sequence divergence between
alternative orientations of a chromosomal inversion polymorphic within An. gambiae and An.
coluzzii (2La and 2L+a) has been noted previously
( 32, 33) but has not been adequately explained.
Other species in the complex are fixed for either
2La (An. arabiensis and An. merus) or 2L+a (An.
quadriannulatus and An. melas) ( 31) (Fig. 5A). In
An. gambiae and An. coluzzii, the 2La arrangement has been shown to confer superior resistance
to desiccating environments ( 34, 35) relative to
2L+a, and its frequency correlates with environmental gradients of aridity ( 10, 32). It has been
argued that 2La introgressed from An. arabiensis
into the presumed 2L+a ancestor of An. gambiae +
An. coluzzii ( 36), a crucial step facilitating the range
expansion of a presumed forest-adapted species
into drier savannas. Sequence divergence-based
estimates of the age of 2La and 2L+a
arrangements relative to the age of the species complex
(Fig. 1C) suggest a radically different scenario in
which 2La/2L+a is an ancient inversion polymorphism that predates the initial diversification of
the entire complex but is maintained as polymorphic only in An. gambiae and An. coluzzii (Fig. 5).
Consistent with this scenario, the topology of the
gene tree built from sequences inside the inversion boundaries indicates that species are grouped
by their 2La or 2L+a karyotype (Fig. 5B). Furthermore, contrary to the longstanding assumption
( 36), our data suggest that 2La introgressed from
ancestral An. gambiae into An. arabiensis, not
the other way around—eventually replacing the
An. arabiensis 2L+a arrangement (Fig. 5A). Our
inference about the direction of 2La introgres-
sion, as well as the underlying phylogenetic hypo-
thesis for the An. gambiae complex, are consistent
with the genome-enabled chromosomal inversion
phylogeny of the An. gambiae complex recon-
structed from ancestral and derived gene orders
at the breakpoints of 10 fixed chromosomal inver-
sions (supplementary text S5 and fig. S27).
Functional insights from differential
We found that introgression mainly involved the
autosomes. Our data suggest that the X chromosome is largely resistant to introgression, consistent
with studies in this and other systems indicating
that the X (or Z) disproportionately harbors factors
responsible for reproductive isolation ( 3, 6, 37–41).
The nature, number, and chromosomal organization of these X-linked factors are unsolved puzzles
for future research, but our data offer one tantalizing clue. An. gambiae males deliver to females
large amounts of 20-hydroxyecdysone (20E) ( 42),
a steroid hormone that increases female fertility
( 43) and fecundity ( 44) and regulates mating behavior and success ( 45). The cytochrome p450 gene
CYP315A1 (AGAP000284) that synthesizes the 20E
precursor ecdysone is located near the distal end
of the X chromosome. Furthermore, this region is
associated with male hybrid sterility between An.
gambiae and An. arabiensis, with the An. gambiae
X chromosome causing inviability in an An.
arabiensis genetic background ( 40). Combined
with our data showing that male An. arabiensis
produce significantly less 20E than An. gambiae
(supplementary text S6 and fig. S28), these observations prompt the hypothesis that divergence in 20E function between the two species
may have a role in speciation through possible
effects on the reproductive fitness of hybrid males.
Pervasive autosomal introgression between
An. arabiensis and the An. gambiae–clade ancestor is consistent with the paucity of sterility
factors across much of the autosomes, although
several autosomal quantitative trait loci have been
mapped in both species ( 40). Accordingly, we
explored the small subset of autosomal genes
(n = 485) that showed no indication of introgression (supplementary text S6), as these are
candidates contributing to reproductive isolation. We found a remarkable overrepresentation
of genes encoding cyclic-nucleotide phosphodiesterases, enzymes that regulate the levels of the
messengers cyclic adenosine monophosphate
and cyclic guanosine monophosphate ( 46), which
in turn control (among other processes) ecdysone synthesis ( 47, 48) (table S16).
Implications for the evolution of
Initial radiation of the An. gambiae complex
was both recent and rapid. Counter to the traditional view ( 49), it is now clear that the ancestor
of the principal malaria vectors An. gambiae and
An. coluzzii separated from other species in the
group approximately 2 million years ago (Ma)
and that An. gambiae and An. coluzzii are distantly related to the other primary vector in the
group, An. arabiensis. Extant populations of
An. gambiae and An. coluzzii are highly anthro-
pophilic vectors, dependent upon humans for
blood meals, adult shelter, and larval breeding
sites, yet anthropogenic influences are unlikely
to have triggered their cladogenesis an estimated
0.5 Ma. Instead, anthropophilic traits are likely
to have developed in conjunction with the ex-
pansion of Neolithic human populations that
occurred more recently. Despite a history of ex-
tensive introgression with An. arabiensis, An.
gambiae and An. coluzzii are behaviorally, phys-
iologically, ecologically, and epidemiologically
distinct from An. arabiensis; the same is true for
An. merus and An. quadriannulatus. Notably,
experimental introgressions of certain autosomal
inversions result in stable heterotic polymor-
phisms, whereas other introgressed autosomal
and X chromosome inversions are rapidly elimi-
nated ( 22), consistent with a role for natural se-
lection in the fate of introgressed regions. Given
evidence that the 2La inversion polymorphism is
maintained by selection in An. gambiae and An.
coluzzii ( 32, 50), it seems likely that its introgres-
sion into An. arabiensis was adaptive, and bidi-
rectional introgressions across the genome between
these species probably contributed to their wide
ecological flexibility and their vectorial capacity.
Broad overlap exists between geographic ranges
of these species, and the potential for ongoing
hybridization and introgression remains, includ-
ing the opportunity for introgression of insecti-
cide resistance alleles ( 51). Our study establishes
a foundation for further study of adaptive intro-
gression in this species complex and its role in
shaping vectorial capacity in this and other ma-
laria vector species complexes.
REFERENCES AND NOTES
1. R. E. Green et al., A draft sequence of the Neandertal genome.
Science 328, 710–722 (2010). doi: 10.1126/science.1188021;
2. K. K. Dasmahapatra et al., Butterfly genome reveals
promiscuous exchange of mimicry adaptations among species.
Nature 487, 94–98 (2012). doi: 10.1038/nature11041;
3. S. H. Martin et al., Genome-wide evidence for speciation with
gene flow in Heliconius butterflies. Genome Res. 23, 1817–1828
(2013). doi: 10.1101/gr.159426.113; pmid: 24045163
4. E. Y. Durand, N. Patterson, D. Reich, M. Slatkin, Testing for
ancient admixture between closely related populations. Mol.
Biol. Evol. 28, 2239–2252 (2011). doi: 10.1093/molbev/
msr048; pmid: 21325092
5. D. Garrigan et al., Genome sequencing reveals complex
speciation in the Drosophila simulans clade. Genome Res. 22,
1499–1511 (2012). doi: 10.1101/gr.130922.111; pmid: 22534282
6. M. A. White, C. Ané, C. N. Dewey, B. R. Larget, B. A. Payseur,
Fine-scale phylogenetic discordance across the house mouse
genome. PLOS Genet. 5, e1000729 (2009).
doi: 10.1371/ journal.pgen.1000729; pmid: 19936022
7. N. Patterson, D. J. Richter, S. Gnerre, E. S. Lander, D. Reich,
Genetic evidence for complex speciation of humans and
chimpanzees. Nature 441, 1103–1108 (2006). doi: 10.1038/
nature04789; pmid: 16710306
8. G. M. Giles, A handbook of the gnats or mosquitoes giving the
anatomy and life history of the Culicidae together with
descriptions of all species noticed up to the present date.
(John Bale, Sons & Danielsson, Ltd., London, 1902).
9. G. Davidson, Anopheles gambiae complex. Nature 196, 907
(1962). doi: 10.1038/196907a0
10. M. Coluzzi, A. Sabatini, V. Petrarca, M. A. Di Deco,
Chromosomal differentiation and adaptation to human
environments in the Anopheles gambiae complex. Trans.
R. Soc. Trop. Med. Hyg. 73, 483–497 (1979).
doi: 10.1016/0035-9203( 79)90036-1; pmid: 394408
11. B. J. White, F. H. Collins, N. J. Besansky, Evolution of Anopheles
gambiae in relation to humans and malaria. Annu. Rev. Ecol.
Evol. Syst. 42, 111–132 (2011). doi: 10.1146/annurev-ecolsys-