Prospects for an Invasion: Competition Between
Aedes albopictus and Native Aedes triseriatus
in which N. represents the initial number of
females, A, the number of females emerging
on day x, and wx the mean winglength of
females emerging on day x;flw,) predicts
numbers of female offspring based on female
body size; and D is an estimate of the delay
between female emergence and first oviposition. Separate studies of individual females, offered blood daily after emergence,
TODD P. LIVDAHL AND MICHELLE S. WILLEY
Competition between larval populations of the native North American treehole
mosquito Aedes triseriatus and Aedes albopictus, recently introduced from Asia to
North America, was assessed by comparing per capita growth rate estimates for
experimental cohorts of larvae developing under a variety of initial density combina- provide estimates of D 11.9 days for A.
tions in fluid obtained from tires or from treeholes. Estimates of carrying capacities triseriatus and D = 14.2 days for A. albopicand competition coefficients indicate that competition between the two species will tus. Linear functions describe size-depenresult in stable coexistence in treehole communities but local extinction of A. dent fecundity: j(wx) = 23.2w. - 51 for
triseriatus in tire habitats.
A. triseriatus (r2 = 0.38) and ftwx) =
17.2wx- 14 (r2 = 0.03) for A. albopictus
EDES ALBOPICTUS, A CONTAINERined competition between A. albopictus and (10).
breeding mosquito, has begun to A. triseriatus, a common filter-feeding and
Regression of r' on the initial densities of
colonize the North American conti- browsing inhabitant of treeholes and tires A. triseriatus and A. albopictus can be exnent during the last several years, after its throughout the eastern United States (5), to pressed in the form
introduction within imported used automo- determine whether A. triseriatus can prevent
bile tires. Successful colonization by this the establishment of this ecologically similar
dNi
i
rm + biNi
(2)
bjNj
potential vector for disease transmission, species. Previous comparisons of separate
most notably dengue fever (1), and which aspects of single and mixed species cohorts
occupies much of Asia and the Pacific re- using artificial media in jars, tires, and
When (rTm + bNi + bNj) is set equal to
gion, was first observed in Houston in tons enable predictions that interactions of zero and Nj is plotted as a function of N.,
1985, where it had already achieved domi- A. albopictus with resident species may result the result is a line that defines the combinance in artificial containers (2). Aedes atin a long-term establishment in tires (6).
nations of initial densities that result in no
bopictus has since become widespread, occurEggs obtained from laboratory colonies population growth for species i, the zero
ring in much of the southeastern and into of A. albopictus and A. triseriatus (7) were growth isocline (Fig. 1). Density combinathe midwestern United States (3). The in- hatched 24 hours before the experiment. tions falling below an isocline permit controduction of A. albopictus has public health
Larvae of each species were disbursed, in all tinued growth of the species associated
significance, and the ecological outcome of density combinations of 7, 14, 28, and 42 with that isocline; density combinations
this introduction will enhance our under- with single species controls for each density, above an isocline will result in the species'
standing of community structure. Successful into uncovered 120-mi plastic cups contain- decline.
establishment of A. albopictus without the ing 0.5 g of dried leaf litter and 100 mI of
In treehole fluid (Fig. 1A), the isoclines
displacement of native filter-feeding and treehole or tire fluid, depending on the intersect at positive densities for both spebrowsing species will imply that evolution- habitat treatment (8). For treehole simula- cies; long-term interaction between these
tions, 25 ml of fluid was replaced weekly species should therefore result in their conary niche diversification in response
prolonged competition between larval stages is with fresh tree drainage water (stemflow), tinued coexistence in treeholes. In tire fluid
not necessary for the coexistence of ecolog- collected with a slit hose caulked about a (Fig. LB), the isoclines do not intersect at a
ically similar species.
maple tree. Both habitat simulations re- positive density for A. triseriatus; long-term
Aedes albopictus has demonstrated an apti- ceived weekly replacement of evaporated association between these species should
tude for rapid colonization of artificial water with distilled water. Experimental result in the competitive exclusion from tires
maintained in
aquatic larval habitats, such as discarded populations
of A. triseriatus by A. albopictus.
insectary
automobile tires (3). Egg diapause respons- (220 + IC, -70% relative humidity, and
The Lotka-Volterra model of two-species
es to photoperiod and freezing tolerance
16 hours light:8 hours dark photoperiod). competition provides a useful starting point
indicate that A. albopictus immigrated from Three replicates were raised for each two- for the analysis of this interaction, in which
temperate
Japanese populations (4) species density combination and five repli- the per capita rate of change for species i is a
preadapted for physical conditions in much cates of all single-species density levels. Pu- direct analog of Eq. 2
of North America. Despite its colonizing pae were removed from experimental
r.m
dNi
rmiay
ability and tolerance of physical conditions, habitats and placed within vials of water
(3)
the eventual success of A. albopictus may topped with inverted vials. Adults were reK, N
Ntdt 'i Ki
hinge on its ability to interact with native moved daily; female wings were measured at
Within this framework, the stability of the
magnification x25.
species within communities under invasion,
Per capita growth rates were estimated for interactions depends on a combination of
particularly the mosquitoes that occupy treeall larval cohorts from the composite index the competition coefficients (a%), which
holes and tires in eastern North America.
This experiment addresses the potential (9)
convert the per capita influence of species j
on the growth rate of species i into the
for competitive success of A. albopwtus as a
colonist of two habitats, tires and waterequivalent number of species i and carrying
InL
Axf(wx)]
filled treeholes, both of which maintain wellcapacities (Ks). These quantities are calculat(1) ed indirectly from the intercept and regresdeveloped native mosquito fauna. We examsion coefficients of Eq. 2.
D +
x
x
The difference in the predicted outcome
of
Clark
Department
Biology,
University, Worcester,
for the two fluids is clarified by comparing
MA 01610.
=
A
=
+
car-
to
were
an
[.XAxf(wx)yiAxf(wx)
12 JULY 1991
REPORTS
189
with A. albopictus at very low densities.
Sufficient time may not have elapsed since
the
introduction of A. albopictus to enable
s
abo
g B|A.
gi '0@
|
80
lsdln.
01 8
the
test
of these predictions in field popula1
tions, but coexistence between A. albopictus
and A. triseriatus within treehole communi10 to 60
oCD0 60 0
01
ties is suggested by the appearance of A.
triseriatus and A. albopictus as the first and
Atdsertafus
~
A tisea~
J1
second most abundant filter-feeding species
0
4
in glass jars attached to trees in a Louisiana
forest
(14), making up 81% and 17% of the
z
z5
filtering and browsing larvae collected. The
co-occurrence of these two species has also
been noted in other studies of treehole
40
0
20
60
80
100
40
20
60
0
80
100
communities
(15). Field studies of A. alNo. of A albopictus per 100 ml
No. of A albopitus per 100 ml
aegypti in artificial habitats,
with
A.
bopictus
Fig. 1. Density combinations that permit equilibrium for Aedes triseriatus and A. albopictus. Hatched
regions overlay combinations for each species for which the 95% confidence interval of the predicted tires in Texas (2) and cemetery vases in
per capita rate includes zero in (A) treehole and (B) tire-simulated habitats. The equilibrium Florida (15), suggest success byA. albopictus
combination of densities is circled for each fluid treatment. Differences between the intercepts of the as a competitor in such containers, with a
two isoclines are statistically significant on both axes in treehole fluid and on the vertical axis in tire fluid concomitant reduction ofA. aegypti popu(P < 0.01 in each case); the difference between intercepts on the horizontal axis in tire fluid is not lations. Aedes triseriatus is only abundant in
significant. Standard errors for these t tests were obtained with the jackknife technique (11).
tire habitats located within shaded areas, and
few detailed census data are available for
K. and ag in each fluid for each species (Fig. significant and opposite differences in a,7's such habitats in areas that have been reached
2). The difference does not result from for the two species. If these coefficients by dispersing A. albopictus. However, the
altered Ks's alone, which are significantly reflect overlap of resource. use, the differ- potential for A. albopictus to dominate artilower in tire fluid for both species, reduced ences suggest that resources exist within ficial habitats has been clear since its discovby nearly the same amount for each. The tires that can be used much more effectively ery in Texas, where A. albopictus' abundance
weekly addition of nutrients and removal of by A. albopictus than A. triseriatus, and that was three times that of A. aegypti and 20
metabolic wastes by replacing 25 ml with treehole fluid contains resources that each times that of A. triseriatus (2).
Although A. triseriatus may be the most
stemflow in treehole fluid may account for species exploits differentially. Detailed studies of the larval diets of these two species in likely potential treehole competitor, other
much of this difference in Ki.
The differential competitive outcome in the two habitat types should be informative. treehole species might interact with A. alThis analysis could be enhanced by the bopictus. Differences among filter-feeding
the two fluid types results primarily from
incorporation of a wider variety of initial species in vulnerability to the predatory
densities of the two species, which would mosquito larva Toxorhynchites rutilus in the
60 r
permit tests for nonlinear concave responses Southeast could influence the outcome of
to density, known to occur in A. triseriatus theA. albopictus invasion. Additional species
50sh
growing in an artificial medium (11), as well of potential importance to the establishment
Ka-as nonadditive interactions between the of A. albopictus in treehole communities
K
+ 3.19m
However, the narrow isocline include Anopheles barberi, a potential filtercompetitors.
40
regions defined by the linear model for A. feeding competitor that becomes a facultaKt
triseriatus in treehole fluid suggest that non- tive predator on small larvae during its
2.73'
30 L
linear isoclines resulting from such nonlin- fourth instar (16), and three additional filear models should not vary substantially ter-feeding and browsing species: Aedes
1.0
from this linear case, leaving intact the pre- hendersoni, Orthopodomyia signfera, and 0.
-2.29*
diction of coexistence. More likely, our pre- alba. However, all four of these potential
_
_ 1..
i _
0.6
diction
of competitive exclusion of A. trise- competitors have marked preferences for
a
riatus in tire fluid may require modification restricted types of treeholes (5), and are
0.4
CL at
4.88*
to include the possibility of its coexistence therefore unlikely to present consistent ob02
J1
J1
q
J1
1-.
-
Tire
Trehoie
Fluid ource
Fig. 2. Comparisons of parameters of competiantd K. depict
tion in treehole and tire fluid.
carrying capacities forA. triseriatus an dA. albopictus. Competition coefficients translate the competitive effect of an A. albopictus indiviidual on A.
tnseriatus into the equivalent number ofA. triseriatus individuals (ama), and the revers(e translation
(acL). Standard errors for each estimaLte, obtained
with the jackknife technique (11), mre shown as
vertical bars. The t statistics for diifferences of
parameters estimated for treehole arnd tire fluid
treatments are shown (*P < 0.05, * *P < 0.01).
190
Table 1. Responses of per capita growth rate estimates for each species to the initial density of
Aedes triseriatus and A. albopictus larval cohorts in two types of fluid. Regression analyses are based
on Eq. 2. Standard errors for each regression term and the coefficient of determination (r2) are
shown. All regression coefficients are statistically significant (P < 10`-). Units are as follows:
rm, days'-; b1 and b2, days-' x (individuals/100 ml)-1. The numbers of r' values obtained for
each regression (n) differ because some cohorts produced no survivors, especially A. triseriatus
growing at high densities in tire fluid (10).
Fluid
Species
n
rm
b2
bi
0.0029 -0.0012 ± 0.0001 -0.0005 ± 0.0001 0.74
Treehole triseriatus 65 0.0514 +
albopictus 64 0.0798 ± 0.0036 -0.0015
triseriatus 35 0.0591 + 0.0070 -0.0018
Tire
albopwtus 65 0.0904 + 0.0036 -0.0020
+
+
+
0.0001 -0.0011
0.0003 -0.0015
0.0001 -0.0005
+
+
+
0.0001 0.80
0.0003 0.66
0.0001 0.82
SCIENCE, VOL. 253
stacles to the establishment of A. albopictus;
a large fraction of treeholes contain only A.
triseriatus as potential competitors (5). Our
results indicate that competition with the
predominant species in treehole communities, A. triseriatus, will be insufficient to
prevent the establishment of A. albopictus.
1.
2.
3.
4.
5.
REFERENCES AND NOTES
D. ShroyerJ. Am. Mosq. Control 2, 424 (1986).
D. Sprenger and T. Wuithiranyagool, ibid., p. 217.
Am. Mosq. Control Assoc. Newsl. 16 (no. 3), 13
(1990); ibid. 16 (no. 2), p. 5; ibid. 15 (no. 1), 15
(1989).
W. Hawley, P. Reiter, R. Copeland, C. Pumpuni,
G. Craig, Science 236, 1114 (1987).
W. Bradshaw and C. Holzapfel, Oecologia 57, 239
(1983); ibid. 74, 507 (1988); R. Copeland and G.
Craig, Ann. Entmol. Soc. Am. 83, 1063 (1990).
6. W. Black, K. Rai, B. Turco, D. Arroyo, J. Med.
Entomol. 26, 260 (1989); C. Ho, A. Ewert, L.
Chew, ibid., p. 615.
7. Aedes albopictus and one A. triseriatus population
were established by mosquitoes collected in New
Orleans, LA. Additional A. triseriatus populations
were established by females from central Massachusetts. A three-way analysis of variance of the interactions of Louisiana and Massachusetts A. triseriatus
with A. albopictus reveal no statistically significant
differences betweenA. triseriatus strains (T. Livdahl,
unpublished data), and the results presented here
combine the results of both strains.
8. After removing mosquito larvae, without removing
detritus, fluid contents of ten treeholes and ten tires
were pooled and homogenized with a canoe paddle
during fluid disbursement.
9. T. Livdahl and G. Sugihara,J. Anim. Ecol. 53, 573
(1984).
10. Some cohorts failed to produce adult females. In
such cases, the cohort was pooled with emergence
data from another cohort within the same experimental cells, with r' calculation based on the total
number of individuals for the pooled cohorts.
11. R. Sokal and F. Rohif, Biometry (Freeman, New
York, 1981), pp. 795-799.
12. J. Hard, W. Bradshaw, D. Malarkey, Oikos 54, 137
(1989).
13. E. Walker and R. Merritt, Environ. Entomol. 17, 199
(1988).
14. E. Schreiber, C. Meek, M. Yates, J. Am. Mosq.
Control 4, 9 (1988).
15. G. O'Meara, "The spread of Aedes albopictus and
Aedes bahamensis in Florida," paper presented at the
1990 Annual Meeting of the Entomological Society
of America, New Orleans, 2 to 6 December 1990.
16. J. Peterson, H. Chapman, 0. Willis, Mosq. News 29,
134 (1969).
17. We thank J. Edgerly-Rooks, S. Gaimari, and J.
Turner for technical assistance, 0. Sholes, R. Bertin,
and two anonymous reviewers for helpful suggestions, J. Freier for A. albopictus eggs, and NIH for
financial support (R15AI27940 to T.P.L.).
21 January 1991; accepted 30 April 1990
bility, which allows them to serve as nucleation sites for proper folding oflarger RNAs
(5). In this report we identify the structural
properties that give rise to the sequence
HANs A. HEUS* AND ARTHUR PARDIt
requirements and high stability of hairpins
containing GNRA loops.
The GCAA and GAAA loops were chosen
The most frequently occurring RNA hairpins in 16S and 23S ribosomal RNA contain
a tetranucleotide loop that has a GNRA consensus sequence. The solution structures for these nuclear magnetic resonance
of the GCAA and GAAA hairpins have been determined by nuclear magnetic (NMR) studies because they are the most
resonance spectroscopy. Both loops contain an unusual G-A base pair between the first frequently occurring members of the GNRA
and last residue in the loop, a hydrogen bond between a G base and a phosphate, family. The full sequences of the two hairextensive base stacking, and a hydrogen bond between a sugar 2'-end OH and a base. pins are shown in Fig. 1 (6). Proton and 31p
These interactions explain the high stability of these hairpins and the sequence resonances were assigned by standard tworequirements for the variant and invariant nucleotides in the GNRA tetranucleotide dimensional (2-D) NMR techniques (7).
Short (<4.5 A) 'H-'H distances were idenloop family.
tified by 2-D nuclear Overhauser effect
R
NA HAmRPINs CONSISTING OF A either G or A) (1). The GNRA hairpin loops (NOE) spectroscopy, and Fig. 2 shows a
double-stranded stem and a single- also occur with high frequency in catalytic, portion of the spectrum that was used to
stranded loop are among the most phage, and eukaryotic signal-recognition assign aromatic and sugar C-i' proton resonances for the GCAA hairpin. Assignment
common structural motifs in RNA. RNA particle (SRP) RNAs (2). The solution
hairpin loops have been viewed simply as a structure of another frequently occurring
RNA hairpin, which belongs to the UNCG
means for the RNA strand to fold back on
9
G3t
itself, and it was thought that the sequence family of tetranucleotide loops, has been
G29
of the loop might not be very critical. How- reported (3).
G5
1
RNA hairpins containing the GNRA
ever, recent phylogenetic and thermody(1
.p Q
namic studies have revealed strong size and loop sequence have been shown to be subU15
sequence dependencies for hairpin loops in stantially more stable, with melting transi0 E
U121 C6
.p
RNA (1). In particular, hairpins containing tion temperatures (Tm) more than 40C
G9
CL
higher than other less frequently occurring
a tetranucleotide loop seem to be very imc
Il, ci
#C
.c
portant. Phylogenetic studies show that in sequences (4). These thermodynamic and
ribosomal RNAs (rRNAs) tetranucleotide phylogenetic data led to the proposal that
loops comprise more than 55% of all loops, GNRA tetranucleotide loops may be evoluand more than 50% of all tetranudeotide tionarily selected because of their high sta- 1W,
F. A14..,5.8 .15.6 5.4 5.2 3.8 3.6
loops have the consensus sequence GNRA
6.0
en, (ppm)
(where N can be any nucleotide and R is
A A
CA
Structural Features That Give Rise to the Unusual
Stability of RNA Hairpins Containing GNRA Loops
0
!GoS lo°
5G
C-G
5G
A
C-G
G-Cl 0
G-Cl 0
Department of Chemistry and Biochemistry, University
of Colorado at Boulder, Boulder, CO 80309.
*Present address: Department of Biophysical Chemistry,
University of Nijmegen, Tournooiveld, 6525 ED Nij-
megan, Netherlands.
tTo whom correspondence should be addressed.
12 JULY 1991
G-C
1 G-U
G-C
1 G-U
Ppp
A
Pp
UA
OH
UA
OH
Fig. 1. The full sequences for the GCAA and
GAAA RNA hairpins (17).
a .-.
aDL
Fig. 2. A portion of the 2-D NOE spectrum of
the GCAA hairpin in D20 showing the aromatic
H2/H8/H6 to sugar HI/H5 region. This spectrum was taken with -1.9 mM RNA hairpin in
100 mM NaCI, 10 mM sodium phosphate, pD =
6.8, at 25°C. Examples of sequential resonance
assignment pathways involving H6/H8(n)-
Hi'9(n)-H6/H8(n+1) resonances are illustrated.
Note the unusual upfield shift of the G9
H-i'.
REPORTS
191