Syst. Appl. Microbiol. 11, 128 (1989); J. A. Eisen, J. Mol.
Evol. 41, 1105 (1995).
13. Searches were done as in (7).
14. A. C. Ferreira et al., Int. J. Syst. Bacteriol. 47, 939
(1997).
15. The distribution of the 64 trinucleotides was computed
for the megaplasmid and the two chromosomes. The
compositions of the megaplasmid and chromosome II
were compared with the x2 test, with each other, and
with similarly sized samples of chromosome I chosen
uniformly at random. These tests indicate that all three
elements have significantly different trinucleotide com-
position (P , 0.01). Although the size of the plasmid is
too small to obtain statistically meaningful compari-
sons of trinucleotide composition, its GC content is
significantly different from the remaining genome
(56% compared with 67%) ( Tables 1 to 5).
16. J. G. Lawrence and H. Ochman, J. Mol. Evol. 44, 383
(1997).
17. S. Tirgari and B. E. B. Moseley, J. Gen. Microbiol. 119,
287 (1980).
18. A. Tanaka, H. Hirano, M. Kikuchi, S. Kitayama, H.
Watanabe, Radiat. Environ. Biophys. 35, 95 (1996).
19. L. M. Markillie, S. M. Varnum, P. Hradecky, K. K. Wong,
J. Bacteriol. 181, 666 (1999).
20. J. K. Setlow and D. E. Duggan, Biochim. Biophys. Acta
87, 664 (1964).
21. M. J. Daly, L. Ouyang, P. Fuchs, K. W. Minton, J.
Bacteriol. 176, 3508 (1994).
22. B. E. Moseley and D. M. Evans, J. Gen. Microbiol. 129,
2437 (1983); P. R. Tempest and B. E. Moseley, Mol.
Gen. Genet. 179, 191 (1980).
23. J. A. Eisen, thesis, Stanford University, Stanford, CA
(1999).
24. C. Bauche and J. Laval, J. Bacteriol. 181, 262 (1999).
25. M. J. Bessman, D. N. Frick, S. F. O?Handley, J. Biol.
Chem. 271, 25,059 (1996).
26. D. M. Sweet and B. E. Moseley, Mutat. Res. 34, 175
(1976).
27. A gene is considered recently duplicated if its great-
est degree of similarity is to another gene in the D.
radiodurans genome relative to genes from other
completed genomes.
28. K. S. Udupa, P. A. O?Cain, V. Mattimore, J. R. Battista,
J. Bacteriol. 176, 7439 (1994).
29. An electrophoretic mobility-shift assay identified a
DNA binding activity from soluble cell extracts spe-
cific for the D. radiodurans genomic repeat. A trun-
cated DNA ligand tracking an inverted repeat sug-
gests that the protein responsible for binding acts as
a dimer. Binding was demonstrated in recA2 cells.
30. M. E. Boling and J. K. Setlow, Biochim. Biophys. Acta
123, 26 (1966).
31. H. J. Agostini, J. D. Carroll, K. W. Minton, J. Bacteriol.
178, 6759 (1996).
32. K. Furuya and C. R. Hutchinson, FEMS Microbiol. Lett.
168, 243 (1998).
33. The predicted molecular weight and isoelectric point
are consistent with a protein that increases in abun-
dance after irradiation (18). This protein is likely the
nuclease reported by R. E. Mitchel [Biochim. Biophys.
Acta 621, 138 (1980)].
34. C. S. Lin, H. C. Wang, T. Y. Wong, J. K. Liu, Biochem.
Mol. Biol. Int. 45, 651 (1998).
35. M. Sandigursky and W. A. Franklin, Curr. Biol. 9, 531
(1999).
36. C. Mun, J. Del Rowe, M. Sandigursky, K. W. Minton,
W. A. Franklin, Radiat. Res. 138, 282 (1994).
37. G. J. Sharples and D. R. F. Leach, Mol. Microbiol. 17,
1215 (1995).
38. S. Kitayama, M. Kohoroku, A. Takagi, H. Itoh, Mutat.
Res. 385, 151 (1997).
39. P. D. Gutman, P. Fuchs, L. Ouyang, K. W. Minton, J.
Bacteriol. 175, 3581 (1993).
40. Supported by the U.S. Department of Energy, Of-
fice of Biological and Environmental Research, Co-
operative Agreement DE-FC02-95ER61962. We
thank E. V. Koonin for thoughtful contributions to
this manuscript. Correspondence and requests for
material should be addressed to C.M.F. (E-mail:
drdb@tigr.org). The sequences have been deposited
in GenBank with accession numbers AE00513,
AE001825, AE001826, and AE001827 for chromo-
some I, chromosome II, the megaplasmid, and the
plasmid, respectively.
3 August 1999; accepted 22 October 1999
Species Diversity and Invasion
Resistance in a Marine
Ecosystem
John J. Stachowicz,1* Robert B. Whitlatch,1 Richard W. Osman2
Theory predicts that systems that are more diverse should be more resistant
to exotic species, but experimental tests are needed to verify this. In experi-
mental communities of sessile marine invertebrates, increased species richness
significantly decreased invasion success, apparently because species-rich com-
munities more completely and efficiently used available space, the limiting
resource in this system. Declining biodiversity thus facilitates invasion in this
system, potentially accelerating the loss of biodiversity and the homogeniza-
tion of the world?s biota.
Along with habitat modification, the intentional
or accidental introduction of new species by
humans is a leading cause of the global biodi-
versity crisis (1). Because biological invasions
can dramatically alter community composition
and ecosystem function (2?4) and cause con-
siderable economic damage (5), there is sub-
stantial interest in understanding why and how
successful invasions occur. Although all sys-
tems do not appear to be equally invasible (3, 6,
7), factors determining the susceptibility of a
community to invasion remain unclear. Theory
predicts that species-rich communities should
be less susceptible to invasion because of a
more complete utilization of resources (6, 8, 9),
but data in support of this prediction have been
elusive (10). Some observational studies do
support a positive relation between biodiversity
and invasion resistance (7, 8), but others do not
(11). However, the large number of uncon-
trolled factors in these studies makes interpret-
ing these findings difficult; manipulative exper-
iments are needed to assess the effect of species
richness on invasion success more directly.
Studies of terrestrial grasslands and aquatic mi-
crobial communities in laboratory microcosms
have demonstrated that species-rich communi-
ties are more resistant to being invaded by
additional species than are species-poor com-
munities (12). However, no studies have inves-
tigated this relation by using exotic species that
currently pose an invasion threat to natural sys-
tems, and few studies offer evidence for the
mechanisms underlying these patterns.
A growing number of marine invertebrates
have been introduced to the coastal waters of
southern New England (13). In some habitats,
these species have invaded successfully and
reduced the abundance of native species,
whereas in others, they have been unsuccessful
and the native community remains unchanged
(14). Some of these invaders have become
locally dominant space holders, including the
colonial ascidian Botrylloides diegensis, native
to the Pacific Ocean (13). By introducing inva-
sive species recruits (,1 week old) into exper-
imentally assembled epifaunal communities
with varying numbers of native species, we
tested the effects of native community species
richness on the ability of Botrylloides to invade
coastal habitats.
Experimental communities were composed
of zero to four native species (Fig. 1). This
range of diversity treatments was selected be-
cause most of the space in undisturbed areas
that were equal in size to our communities (100
cm2) was occupied by three to four species.
Each community consisted of 25 2-cm-by-2-cm
tiles that fit on tracks bolted to a larger substrate
(10 cm by 10 cm). Native sessile invertebrates
were cultured on tiles in the field by allowing
individuals to settle on tiles, then these tiles
were ?gardened? weekly to remove all other
species except the target species. Once a single
tile was covered by an individual or colony of a
native species, a number of such tiles were
arranged to produce communities with the de-
sired species richness (Fig. 1). Five Botryl-
loides recruits were interspersed throughout
each community so that there was only one
individual in each row and column of the five-
by-five grid of tiles that composed the commu-
nity (Fig. 1). The remaining 20 tiles in each
community were covered by native species. In
multispecies communities, available space was
divided equally among native species, and the
1Department of Marine Sciences, University of Con-
necticut, 1084 Shennecossett Road, Groton, CT
06340, USA. 2Academy of Natural Sciences, Estuarine
Research Center, St. Leonard, MD 20685, USA.
*To whom correspondence should be addressed. E-
mail: jstach@uconnvm.uconn.edu
R E P O R T S
www.sciencemag.org SCIENCE VOL 286 19 NOVEMBER 1999 1577
spatial arrangement of individuals of different
native species was determined with a random-
number generator. Native species were drawn
at random from among five of the most com-
mon species at the location where experiments
were conducted: the blue mussel (Mytilus edu-
lis), two solitary ascidians (Molgula manhatten-
sis and Ciona intestinalis), the colonial ascidian
(Botryllus schlosseri), and the encrusting bryo-
zoan (Cryptosula pallasiana). Each native spe-
cies combination was replicated four times, and
each level of species richness was replicated
with different species combinations.
Experimental communities were deployed
in the field in eastern Long Island Sound near
Groton, Connecticut (15), and they were mon-
itored until all invaders had either been elimi-
nated or had successfully reproduced (indicat-
ing a successful invasion). Reproductive status
was assessed by the presence of mature brood-
ed larvae visible through the translucent orange
tunic of Botrylloides. Communities were pho-
tographed weekly to evaluate the survival and
reproductive status of invaders and to measure
(using image analysis) the availability of prima-
ry space.
We found decreased survival of Botryl-
loides recruits in communities with higher spe-
cies richness (Fig. 2A). Species richness manip-
ulations explained 73% of the variance in the
survival of invaders (r 5 20.855 and P ,
0.0001). This result was not attributable to
dominant effects of any one species, because
species that were best at resisting invasion in
monospecific communities were not necessari-
ly members of the most resistant multispecies
communities. Conversely, species that were
more susceptible to being invaded when grown
alone (for instance, the slow-growing bryozoan
Cryptosula pallasiana) were often members of
highly resistant multispecies communities. Bot-
rylloides has no obvious unique attributes and
seems likely to be representative of other inva-
sive colonial marine invertebrates in its re-
sponse to native community diversity.
Although we cannot be sure of the mecha-
nism by which increased species diversity en-
hances the resistance of our experimental com-
munities to invasion, our results support the
hypothesis that reduced resource availability is
responsible for the decreased success of inva-
sions in communities with increased diversity
(8, 9). Primary space is often the limiting re-
source in marine hard-substrate communities
(16?18), and repeated measures analysis of
variance indicated significant effects of species
richness (F 5 86.9 and P , 0.0001) and time
(F 5 102.2 and P , 0.0001) on the amount of
open space in each community, as well as an
interaction between these factors (F 5 22.0 and
P , 0.0001). These results are reflective of the
fact that, although there was no initial differ-
ence in the availability of space among com-
munities with species richnesses of one through
four, as the communities developed, more
space became available in communities with
fewer native species (Figs. 2B and 3).
These differences arose because natural
population cycles created large increases in
open space in simple communities but did not
create increases in those that were more com-
plex (Figs. 2B and 3). For example, when the
solitary ascidian Molgula manhattensis was the
only species in our experimental community,
increasing community biomass caused by gre-
garious settlement of conspecifics on the tunics
of adults caused the entire aggregation to
slough off the substrate. This left considerable
Fig. 1. Plan view of experimental communities
showing random arrangement of native species
and interspersion of invasive species recruits
(zero-species treatments not shown). Four repli-
cates of each species combination were assem-
bled, and several different species combinations
were assembled for each level of species richness.
All communities (except those with zero species)
begin with about the same initial cover of native
species.
Fig. 2. (A) Survival of recruits of the exotic
ascidian Botrylloides diegensis and (B) availabil-
ity of free space, which are plotted versus
native community species richness. Availability
of free space is taken as the highest measured
open space that occurred after communities
were fully developed (after 14 days) (see Fig. 3)
and thus represents the amount of resources
(space) freed because of population cycles in
each community. Each dot represents results
(survival of recruits or availability of space)
averaged over the four replicates of a given
species combination, and larger dots indicate
two coincident data points. Statistical analysis
was by linear least squares regression (solid
lines) [in (A), r 5 20.855 and P , 0.0001; in
(B), r 5 20.867 and P , 0.0001]. Results are
qualitatively unchanged when the zero-species
treatment is deleted from the analysis [in (A), r
5 20.839 and P 5 0.0003; in (B), r 5 20.702
and P 5 0.009].
Fig. 3. Time series of the
availability of open space
(symbols represent the
mean; error bars repre-
sent 61 SE) in represen-
tative communities of
each species richness lev-
el (zero species; one spe-
cies, Molgula manhatten-
sis; two species, Molgula
and Botryllus schlosseri;
three species, Molgula,
Botryllus, and Cryptosula
pallasiana; four species,
Molgula, Botryllus, Cryp-
tosula, and Ciona intesti-
nalis). These five commu-
nities are displayed be-
cause they were deployed
at the same time of year and the one-, two-, and three-species treatments are composed of a
subset of the species used in the four-species experiment.
R E P O R T S
19 NOVEMBER 1999 VOL 286 SCIENCE www.sciencemag.org1578
bare space (;70% of total) (Fig. 3), allowing
Botrylloides recruits that had survived beneath
the Molgula ?canopy? to expand. A similar
pattern was observed among communities ini-
tially composed of only the semelparous colo-
nial ascidian Botryllus schlosseri (19) (not
shown), which died after reproducing. ?Boom
and bust? population cycles are characteristic of
many epifaunal marine invertebrates (17?20),
and this feature may contribute substantially to
the susceptibility of these simple communities
to invasion.
Species-rich communities appear to be buff-
ered from such fluctuations in space availability.
Although the abundance of each species in mul-
tispecies communities varied, these variations
were out of phase, and the amount of open space
that became available was less than that in sim-
pler communities. For example, communities
that initially contained Botryllus, Molgula,
Cryptosula, and Ciona showed little change in
the availability of free space throughout the
course of the experiment (Fig. 3). This resulted
from the sequential replacement of species;
when one species declined, others increased
and thus maintained high total cover. Be-
cause little space became available in these
experiments, Botrylloides recruits had little
room to grow, and few survived (Fig. 2).
The consistently high cover in more spe-
cies-rich communities may also reduce new
recruitment of exotic invaders into these
communities because these organisms typi-
cally do not settle directly on resident adults
(21). In theory, a single species that could
effectively monopolize space for a long period
of time could resist invasion at least as well as a
multispecies assemblage. However, the exis-
tence of such species in shallow-water epifaunal
communities such as these appears unlikely be-
cause of the short life-span of most species, the
absence of a rigid competitive hierarchy, and the
importance of ?priority? effects (17). Differenc-
es in primary space availability appear to
drive the relation between diversity and inva-
sibility in this system, but this model should
be applicable to any system in which the
limiting resources (such as light or nutrients)
are clearly identifiable.
All native species used in our experiments
could be placed in the same trophic and func-
tional groupings, as they are all sessile, suspen-
sion-feeding invertebrates. As has been pro-
posed for other communities (22), functionally
redundant species may represent a form of bi-
ological insurance against the inevitable loss of
any one species as a consequence of natural
population cycles or disturbances. Our results
lend empirical support to this idea from the
marine environment, strengthening the argu-
ment for efforts to preserve naturally occurring
biodiversity, regardless of whether some spe-
cies are functionally similar. Because biodiver-
sity loss promotes invasion and successful in-
vasion may further decrease biodiversity (2?4),
a negative feedback cycle may be initiated that
ultimately results in severe impoverishment and
homogenization of the global biota.
References and Notes
1. D. S. Wilcove, D. Rothstein, J. Dubow, A. Philips, E.
Losos, Bioscience 48, 607 (1998).
2. J. T. Carlton and J. B. Geller, Science 261, 78 (1993).
3. J. T. Carlton, in San Francisco Bay: The Urbanized
Estuary, T. J. Conomos, Ed. (California Academy of
Sciences, San Francisco, 1979), pp. 427?444.
4. F. H. Nichols, J. K. Thompson, L. E. Schemel, Mar. Ecol.
Prog. Ser. 66, 95 (1990); J. Travis, Science 262, 1366
(1993).
5. Harmful Non-Indigenous Species in the United States
(Office of Technology Assessment, Washington, DC,
1993).
6. M. D. Fox and B. J. Fox, in Ecology of Biological
Invasions: An Australian Perspective, R. H. Groves and
J. J. Burdon, Eds. (Australian Academy of Sciences,
Canberra, 1986), pp. 57?66.
7. M. B. Usher, Biol. Conserv. 44, 119 (1988); A. N.
Cohen and J. T. Carlton, Science 279, 555 (1998).
8. C. S. Elton, The Ecology of Invasions by Animals and
Plants (Methuen, London, 1958).
9. T. J. Case, Proc. Natl. Acad. Sci. U.S.A. 87, 9610
(1990).
10. G. R. Robinson, J. F. Quinn, M. L. Stanton, Ecology 76,
786 (1995); J. V. Robinson and J. E. Dickerson, Oecologia
61, 169 (1984); S. L. Pimm, in Biological Invasions: A
Global Perspective, J. A. Drake et al., Eds. (Wiley, New
York, 1989), pp. 351?367; D. Simberloff, in Ecology of
Biological Invasions of North America and Hawaii, H. A.
Mooney and J. A. Drake, Eds. (Springer-Verlag, New
York, 1986), pp. 3?26.
11. A. Planty-Tabacchi, E. Tabacchi, R. J. Naiman, C. De-
ferrari, H. Decamps, Conserv. Biol. 10, 598 (1996);
S. K. Wiser, R. B. Allen, P. W. Clinton, K. H. Platt,
Ecology 79, 2071 (1998); T. J. Stohlgren et al., Ecol.
Monogr. 69, 47 (1999).
12. J. McGrady-Steed, P. M. Haris, P. J. Morin, Nature 390,
162 (1997); D. Tilman, Ecology 78, 81 (1997); J. M. H.
Knops et al. Ecol. Lett. 2, 286 (1999).
13. J. Carlton, Conserv. Biol. 3, 265 (1989); J. Berman, L.
Harris, W. Lambert, M. Buttrick, M. Dufresne, Conserv.
Biol. 6, 432 (1992).
14. R. W. Osman and R. B. Whitlatch, Mar. Ecol. Prog.
Ser. 117, 111 (1995).
15. Experimental communities were suspended from
floating docks so that they were 1.0 m beneath the
water surface. The docks naturally support a lush
growth of epifaunal invertebrates dominated by the
species used in the experiments.
16. P. K. Dayton, Ecol. Monogr. 41, 351 (1971).
17. J. P. Sutherland and R. H. Karlson, Ecol. Monogr. 47,
425 (1977); L. W. Buss and J. B. C. Jackson, Am. Nat.
113, 223 (1979).
18. R. W. Osman, Ecol. Monogr. 47, 37 (1977).
19. R. K. Grosberg, Evolution 42, 900 (1988).
20. K. D. McDougall, Ecol. Monogr. 13, 321 (1943).
21. R. W. Osman and R. B. Whitlatch, J. Exp. Mar. Biol.
Ecol. 190, 199 (1995).
22. S. Naeem and S. Li, Nature 390, 507 (1997); D.
Tilman et al., Science 277, 1300 (1997); D. Tilman
and J. A. Downing, Nature 367, 363 (1994).
23. We thank M. Berger, H. Lisitano, E. Rogers, and S.
Smith for assistance conducting these experiments
and the Jesse B. Cox Charitable Trust and NSF for
providing funding for our work. A. Lohrer, P. Renaud,
and two anonymous reviewers provided helpful com-
ments on earlier drafts of the manuscript.
29 June 1999; accepted 11 October 1999
Structural Analysis of the
Mechanism of Adenovirus
Binding to Its Human Cellular
Receptor, CAR
Maria C. Bewley, Karen Springer, Yian-Biao Zhang,
Paul Freimuth,* John M. Flanagan*
Binding of virus particles to specific host cell surface receptors is known to be
an obligatory step in infection even though the molecular basis for these
interactions is not well characterized. The crystal structure of the adenovirus
fiber knob domain in complex with domain I of its human cellular receptor,
coxsackie and adenovirus receptor (CAR), is presented here. Surface-exposed
loops on knob contact one face of CAR, forming a high-affinity complex.
Topology mismatches between interacting surfaces create interfacial solvent-
filled cavities and channels that may be targets for antiviral drug therapy. The
structure identifies key determinants of binding specificity, which may suggest
ways to modify the tropism of adenovirus-based gene therapy vectors.
Many viral infections are initiated by the
specific binding of specialized proteins or
attachment factors on the virion?s surface to
glycoprotein receptors on the surface of host
cells. Enveloped viruses, such as human im-
munodeficiency virus (HIV), attach to host
cells by means of spike-like membrane gly-
coproteins, whereas most nonenveloped vi-
ruses, such as poliovirus, attach by means of
specialized domains integral to their capsids.
Adenoviruses (Ad) are hybrids: They are
nonenveloped but have trimeric fibers (320 to
587 residues) emanating from the vertices of
their icosahedral capsid, which terminate in
Biology Department, Brookhaven National Laborato-
ry, Upton, NY 11973, USA.
*To whom correspondence should be addressed. E-
mail: flanagan@bnlbio.bio.bnl.gov and Freimuth@bnl.
gov
R E P O R T S
www.sciencemag.org SCIENCE VOL 286 19 NOVEMBER 1999 1579