university of zurich - uzh · 2011. 1. 17. · 1!"!what traits are carried on mobile genetic...
TRANSCRIPT
University of ZurichZurich Open Repository and Archive
Winterthurerstr. 190
CH-8057 Zurich
http://www.zora.uzh.ch
Year: 2011
What traits are carried on mobile genetic elements, and why?
Rankin, D J; Rocha, E P C; Brown, S P
Rankin, D J; Rocha, E P C; Brown, S P (2010). What traits are carried on mobile genetic elements, and why?Heredity, 106(1):1-10.Postprint available at:http://www.zora.uzh.ch
Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch
Originally published at:Heredity 2010, 106(1):1-10.
Rankin, D J; Rocha, E P C; Brown, S P (2011). What traits are carried on mobile genetic elements, and why?Heredity, 106(1):1-10.Postprint available at:http://www.zora.uzh.ch
Posted at the Zurich Open Repository and Archive, University of Zurich.http://www.zora.uzh.ch
Originally published at:Heredity 2011, 106(1):1-10.
1!
What Traits Are Carried On Mobile Genetic Elements, And "!
Why? #!
Daniel J. Rankin1,2, Eduardo P.C. Rocha3 & Sam P. Brown4 $!
1) Department of Biochemistry, University of Zürich, Building Y27, %!
Winterthurstrasse 190,CH-8057 Zürich, Switzerland &!
2) Swiss Institute of Bioinformatics, Quartier Sorge Bâtiment Génopode, CH‑1015 '!
Lausanne, Switzerland (!
3) Institut Pasteur, Microbial Evolutionary Genomics, CNRS, URA2171, F-75015 )!
Paris, France *!
4) University of Oxford, Department of Zoology, South Parks Road, Oxford OX1 "+!
3PS, UK ""!
E-mail addresses: DJR: [email protected], ER: [email protected], SPB: "#!
Corresponding author: Daniel J. Rankin, Department of Biochemistry, University of "%!
Zürich, Building Y27, Winterthurstrasse 190,CH-8057 Zürich, Switzerland. "&!
Telephone: +41 78 648 9905. "'!
Key-words: Mobile genetic elements, toxin-antitoxin, plasmid addiction, mutualisms, "(!
social evolution, microbial ecology ")!
Running title: Evolution of horizontally-transferred genes "*!
Word count: Main text: 5,568 Abstract: 153 #+!
#"!
2!
"!Abstract #!
Whilst like any other organism, prokaryotes can transfer genes vertically, from $!
mother to daughter cell, they can also exchange certain genes horizontally. Genes can %!
move within and between genomes at fast rates because of mobile genetic elements. &!
While mobile elements are fundamentally self-interested entities, and thus replicate '!
for their own gain, they frequently carry genes beneficial for their hosts and/or the (!
neighbours of their hosts. Many genes which are carried by mobile elements code for )!
traits which are expressed outside of the cell. Such traits are involved in bacterial *!
sociality, such as in the production of public goods, which benefit a cells neighbours, "+!
or the production of bacteriocins, which harm a cells neighbours. Here we review the ""!
patterns that are emerging in the types of genes carried by mobile elements, and "#!
discuss the evolutionary and ecological conditions under which mobile elements "$!
evolve to carry their peculiar mix of parasitic, beneficial and cooperative genes. "%!
"&!
3!
Introduction "!
Whilst in all organisms genes are transferred vertically from parent to offspring, many #!
microorganisms can also transfer genes horizontally, independently of reproduction $!
events, via processes that are known collectively as horizontal gene transfer (HGT). %!
Ever since Joshua Lederberg’s pioneering work, it has been known that bacteria could &!
exchange genes horizontally (Lederberg and Tatum, 1946; Zinder and Lederberg, '!
1952). However, it is only recently that both the significance and the scale of (!
horizontal gene transfer has become apparent (Bordenstein and Reznikoff, 2005; )!
Bushman, 2002; Kurland et al, 2003; Ochman et al, 2000; Sorensen et al, 2005; *!
Thomas and Nielsen, 2005) and has even been described by one party as “biology’s "+!
next revolution” (Goldenfeld and Woese, 2007). ""!
"#!
Prokaryotic genomes are now known to vary in the rates of gene gain and loss. Most "$!
prokaryotes are highly dynamic with high rates of gene gain and loss, whereas some "%!
genomes, typically associated with obligatory endosymbiosis, have stable or shrinking "&!
genomes. Within genomes, genes differ in their propensity to be mobile. In addition to "'!
the core genome (~2000 genes in Escherichia coli which are present in the first 20 "(!
sequenced strains), non-core genes contribute significantly to the overall diversity of ")!
gene repertoires in a species, which together with the core are known as the pan-"*!
genome. In a study of E. coli, for example, these non-core genes made up 90% of the #+!
pan-genome when 20 strains were put together (Touchon et al, 2009). Despite the #"!
widespread attention that horizontal gene transfer has received in recent years (Delsuc ##!
et al, 2005; Frost et al, 2005; Gevers et al, 2005; Goldenfeld and Woese, 2007; #$!
Sorensen et al, 2005), the adaptive significance of horizontally transferred genes #%!
remains to be fully addressed. #&!
4!
"!
There are increasing hints of important separations of functions between the vertically #!
transmitted core genome, which encodes fundamental cellular processes, and the $!
horizontally transmissible accessory genome which encodes for a variety of secondary %!
metabolites conferring resistance to specific toxins or antibiotics or the ability to &!
exploit a specific niche ,-./012!.34!5.236178!#++"9!:;2<.3!"#!$%8!#++*=. The '!
accessory genome contains recently acquired functions, mobile genetic elements, non-(!
expressed genes and genes under particular modes of selection such as diversifying )!
selection, frequency-dependent selection and periodic selection. Despite a large and *!
growing body of research into the molecular mechanisms of HGT, the ecological and "+!
evolutionary forces that drive these basic divisions of mobility and function are poorly ""!
understood (e.g. Slater et al, 2008). "#!
"$!
Horizontal transfer exists largely because some mobile genetic elements (MGE) can "%!
move or be mobilized between microbial cells and carry other traits than the ones "&!
strictly essential for their replication (Frost et al, 2005). They encompass a wide set of "'!
different elements, such as transposable elements, plasmids and bacteriophages. In "(!
prokaryotes, mobile genes can be transmitted horizontally in three main ways: ")!
transformation (where naked DNA is taken up from the local environment Chen and "*!
Dubnau, 2004), transduction (transfer of DNA via a bacteriophage vector) or by #+!
conjugation (transfer of DNA via cell-cell contact, most often mediated by plasmids #"!
Cascales and Christie, 2003; de la Cruz et al, 2009; Thomas and Nielsen, 2005)>!##!
Horizontally transferred genes confer a variety of beneficial and negative effects on #$!
their bacterial hosts. #%!
#&!
5!
The costs of mobile elements differ significantly with the element. Bacteriophages are "!
commonly divided into virulent and temperate. Upon successful invasion, virulent #!
phages kill the cell to release their progeny (although filamentous phages are able to $!
export progeny without cell lysis). Temperate phages may act like virulent phages, but %!
may also integrate in the chromosome and replicate with the bacteria (as a quiescent &!
prophage), and some can reside in the cell as plasmid prophages. In this dormant '!
state, the microbial cell may even express useful traits carried by the phages, yet so (!
long as the phage remains functional, the possibility remains that the phage will exit )!
from the quiescent state, kill the host and transmit horizontally. *!
"+!
The introduction of other mobile elements, such as plasmids, integrative conjugative ""!
elements (ICE), mobilisable islands or transposable elements, does not pose the same "#!
life-or-death dilemma. However, these elements are still costly (Diaz Ricci and "$!
Hernandez, 2000). Element replication requires the synthesis of proteins, RNA and "%!
DNA, which incur in a fitness cost. This cost can be low in general and easily "&!
surmounted by other adaptive or addictive traits. Mobile elements often integrate in "'!
chromosomes thereby potentially disrupting important functions ,?12.@!.34!A/B<.38!"(!
#++%=. Transmission often involves the creation of structures, such as conjugation ")!
pili, that can be costly. These structures can them-selves facilitate the invasion of "*!
phages (Rasched and Oberer, 1986). In the case of elements carried by virulent #+!
phages, transmission results in host death. Mobile elements can also be costly because #"!
of the genes they carry to compete with other genetic elements. For example, F-like ##!
plasmids use exclusion proteins to prevent super-infection by other closely related #$!
plasmids which are very highly expressed (Achtman, 1975). #%!
#&!
6!
Mobile elements may also carry genes coding for adaptive traits. They have thus been "!
described as “agents of open source evolution” (Frost et al, 2005), suggesting that #!
they afford their host access to a vast genetic resource which can then be improved $!
upon and made available to other organisms. Many traits which are beneficial to the %!
host are carried by plasmids, including virulence factors, antibiotic resistance, &!
detoxifying agents and enzymes for secondary metabolism (Martinez, 2009; Martínez, '!
2008; Philppon et al, 2002; Yates et al, 2006). As well as having an effect on the (!
bacterial host, many horizontally transferred genes also code for traits which can )!
affect the fitness of a host’s neighbours. This can be either in a positive way, by *!
producing proteins which can have a beneficial effect on the hosts neighbours "+!
degrading enzymes (Livermore, 1995), or in a negative way, by producing substances ""!
which harm the hosts neighbours, such as bacteriocins (Brown et al, 2006; Dykes and "#!
Hastings, 1999; Riley and Wertz, 2002; van der Ploeg, 2005). "$!
"%!
In spite of their potentially positive effects on host fitness, it is important to consider "&!
that MGEs do not necessarily share the same interest as that of the host genome and "'!
can thus be considered first and foremost as infectious agents: they infect their hosts "(!
much in the same way that parasites infect their host (via contact for plasmids, or via ")!
propagules for phages), and can thus persist in spite of any costs they may impose on "*!
the host. As a result of their distinct transmission mechanisms, the interests of MGEs #+!
may not always coincide with the interests of the host genome as they can employ #"!
different means to ensure they get as many copies of themselves into the next ##!
generation (Orgel and Crick, 1980; Wagner, 2009). Whether they confer beneficial or #$!
detrimental effects on a host will depend on the selective forces acting on both the #%!
mobile element, and the host chromosome. #&!
7!
"!
This review will deal with the evolution of MGEs, in particular the evolutionary #!
explanations for why they harbour the particular balance of traits they carry. While $!
the bulk of this review will focus on plasmids, the general results will help to %!
understand other systems of horizontal gene transfer. We will focus on addressing the &!
conditions under which certain gene classes will be favoured if transmitted '!
horizontally. Mobile elements transmit, and are selected, at a variety of biological (!
scales, from within the genome, to between hosts, and in larger “epidemiological” )!
patches (figure 1). Transmission depends on the underlying ecology of the hosts, and *!
generally increases at higher densities, while other factors, particularly those involved "+!
in bacterial sociality, depend strongly on the spatial structure of the population, which ""!
in turn may be influenced by the transmission rate. As such, it is essential to have a "#!
good understanding of the ecology of their bacterial hosts in order to better "$!
understand why certain genes are more likely to be transmitted horizontally. "%!
"&!
In this review we classify the behaviour of mobile genetic elements by their net "'!
influence on their host cell (from parasitic to mutualistic, e.g. Ferdy and Godelle, "(!
2005) and additionally by their net influence on neighbouring cells (whether they turn ")!
their host cell into ‘helpers’ or ‘harmers’ of neighbouring cells). Hamilton proposed a "*!
way of classifying social behaviours into two types of behaviours, based on their #+!
effect on the fitness of a target individual and that of its neighbour ,-.<67@;38!"*'%=> #"!
This classification is commonly used in social evolution theory, and, in it’s ##!
contemporary form, divides social behaviours into “mutualism”, “selfishness”, #$!
“altruism” and “spite” (West et al, 2007b). We therefore distinguish traits that are #%!
kept within the cell (“private” traits, governing the parasitism-mutualism axis) from #&!
8!
those secreted outside the cell (“public” traits, introducing Hamilton’s social space, "!
which govern the helping-harming axis). (Hamilton, 1964; West et al, 2007b), but for #!
a continuum of behaviours: we divide the types of traits that can be expressed on $!
MGEs into four types of “behaviours”, depending on the effect they have on the host %!
cell, and on neighbouring cells (figure 2). We wish to go through this classification &!
and address the types of genes that fall into these categories, and which selective '!
pressures result in them being carried on mobile genetic elements. (!
)!
Private traits – genomic parasites and mutualists *!
Mobile genetic elements as parasites "+!
As many prokaryotes have very large effective population sizes, any genes conferring ""!
a deleterious effect on their host will be selected against, and purged from the "#!
population. However, many mobile elements do exert costs on their hosts (Diaz Ricci "$!
and Hernandez, 2000; Fox et al, 2008), raising the question as to how costly and "%!
poorly transmissible MGEs can persist within a genome (Bergstrom et al, 2000; Levin "&!
and Stewart, 1977; Lili et al, 2007; Stewart and Levin, 1977). Stewart and Levin "'!
(1977) built one of the first models to look at plasmid persistence, and argued that "(!
plasmids could not persist under low rates of horizontal gene transfer (Stewart and ")!
Levin, 1977). This has provoked debate as to whether horizontal transfer is sufficient "*!
for mobile elements to persist in populations, due to their costs on host fitness (e.g. #+!
Bergstrom et al, 2000; Lili et al, 2007; Slater et al, 2008). The persistence of any #"!
parasite is fundamentally determined by its rate of horizontal transmission (Anderson ##!
and May, 1992), which can be readily measured for plasmids and has been found to #$!
vary over eight orders of magnitude among plasmids of E. coli (Dionisio et al, 2002). #%!
Whereas highly mobile plasmids can readily persist as molecular parasites (Bahl et al, #&!
9!
2007a; Bahl et al, 2007b), the persistence of plasmids with extremely low rates of "!
horizontal transfer cannot be readily explained in this way, given any reduction in #!
vertical transmission imposed by the plasmid (see box 1 – Bergstrom et al, 2000; $!
Dionisio et al, 2002). Rates of horizontal gene transfer have primarily been estimated %!
in the laboratory, and are hard to determine for natural populations (Slater et al, &!
2008), yet the question of how parasitic mobile elements can persist remains an open '!
question and important for understanding the conditions under which certain genes (!
evolve to be transmitted horizontally. Here we focus specifically on the evolution of )!
horizontally transferred genes, as opposed to their persistence (readers wishing to *!
better understand the ecological factors involved in the persistence of plasmids are "+!
directed to a recent paper on the subject by Slater et al, 2008). ""!
"#!
Infectious elements such as plasmids are likely to face a trade-off between horizontal "$!
transfer and vertical transmission, mediated by the costs that they impose on their "%!
hosts (Haft et al, 2009; Turner, 2004; Turner et al, 1998). Any trade-off between "&!
horizontal gene transfer and costs to the host can be viewed as a form of the much-"'!
discussed virulence-transmission trade-off (Alizon et al, 2009). However, unlike "(!
purely horizontally-transmitted parasites, selection to reduce costs to the host is likely ")!
to provide additional benefits to the parasite via enhanced vertical transmission. "*!
Studies on conjugative plasmids demonstrated that the cost of mobile elements #+!
generally decreases under selection (Bouma and Lenski, 1988; Turner et al, 1998): #"!
plasmids which exert less of a cost on their hosts have a higher representation in ##!
daughter cells. However, as an increase in vertical transfer generally comes at a cost #$!
to horizontal gene transfer, there will be an optimum where gene transfer via some #%!
mix of both horizontal and vertical mechanisms is maximized (e.g. Paulsson, 2002). #&!
10!
"!
All mobile elements have mechanisms which allow them to replicate and persist #!
within a genome. For example, Class I transposons (reterotransposons) copy $!
themselves and paste the new copy elsewhere in a genome, while most Class II %!
transposons (DNA transposons) cut themselves out and paste themselves elsewhere in &!
a genome. Both plasmids and phages often come with the genetic architecture needed '!
to replicate and transmit themselves to new hosts, either though conjugation or cell (!
lysis. Mobile elements can be subject to resistance from the rest of the genome by )!
several silencing or degrading mechanisms including methylation, restriction, *!
recognition by CRISPR etc. (e.g. Barrangou et al, 2007; Johnson, 2007; Korona and "+!
Levin, 1993). It is likely that hosts will evolve resistance to mobile genetic elements if ""!
they are too costly to the genome. For example, there is a large variance in the "#!
transmissibility of bacterial plasmids (Dionisio et al, 2002), and selection is likely to "$!
act upon hosts in order to reduce the amount of conjugation, if transfer is costly. Thus, "%!
in many plasmids the expression of the conjugation machinery depends on cell "&!
density (Kozlowicz et al, 2006), or is repressed after some lag upon acquisition of the "'!
plasmid ,C;7D716@312!"#!$%8!"**(=, presumably because by then most recipients cells "(!
already contain the plasmid and the costs of conjugation offset the gains in transfer. ")!
"*!
Gene loss is a common threat to mobile elements, such as through excision (as in the #+!
case of transposable elements) or through segregation (in the case of conjugative #"!
plasmids). Segregation occurs when no plasmids are transferred to a daughter cell ##!
during cell division. Plasmids in particular have evolved mechanisms which allow #$!
them to persist in the face of segregation. Known as plasmid “addiction”, these #%!
complexes come with a gene coding for a toxin, as well as a gene coding for an #&!
11!
antitoxin (Arcus et al, 2005; Buts et al, 2005; Van Melderen and De Bast, 2009). The "!
antitoxin is less stable than the toxin, and is broken down at a faster rate. As the #!
addiction genes are on the plasmid, if the plasmid is lost through segregation, the $!
toxin will remain in the cell (as it breaks down at a slower rate), but without the %!
antitoxin the toxin kills the cell. It is worth noting that a similar principle can be seen &!
to operate in restriction modification systems (Kobayashi, 2001). Additional and '!
complementary hypotheses for toxin-antitoxin evolution exist (Magnuson, 2007). For (!
example, it has also been suggested that addiction complexes have evolved not as a )!
way for plasmids to avoid segregation, but in order to exclude competing plasmids *!
(Cooper and Heinemann, 2000; Cooper and Heinemann, 2005; Mochizuki et al, "+!
2006). This is potentially an effective way for the plasmid to maintain itself within a ""!
population, through competition with other plasmids. However, a plasmid that codes "#!
for an addiction complex exerts a cost on the cell, which means such a plasmid will "$!
have a lower fitness with respect to other plasmids. As the plasmid is lost regardless "%!
of whether it kills the cell or not, the complex cannot be explained by invoking direct "&!
fitness benefits to the plasmid. One way that this can occur is if the dead cell is "'!
replaced by another cell carrying the addiction complex (Rankin and Brown, in "(!
preparation). As such, addiction complexes should be seen as a form of life insurance, ")!
where, in the event of the plasmid being lost through segregation, the addiction genes "*!
gain an advantage by other cells with the addiction complex replacing them with a #+!
greater probability than cells without the complex. Thus, addiction complexes can be #"!
viewed as behaving altruistically (see later section) with respect to other addiction-##!
infected cells. This is supported by models of plasmid addiction, which find that #$!
addiction can only evolve when there is strong spatial structure (Mochizuki et al, #%!
2006). #&!
12!
"!
Consideration of the infectious nature of mobile elements highlights the selection #!
pressures driving MGEs to optimize their own parasite function, balancing gains $!
through transmission (vertical and horizontal) with losses through death of the host. %!
The underlying ecology of their bacterial hosts is likely to play a large role in the &!
evolution of mobile element virulence: if the spatial density of hosts is high, favouring '!
horizontal gene transfer, mobile elements will be selected to increase their horizontal (!
transmission, even at a cost to the host. However, if horizontal transmission is reduced )!
to very low levels, then MGEs can only increase their fitness by coding for traits that *!
enhance vertical transmission, for traits that are beneficial to the host (Ferdy and "+!
Godelle, 2005). Costs of mobile elements are likely to be mitigated by coevolution ""!
with their bacterial hosts. If plasmid-chromosome co-evolution results in the "#!
evolution of reduced plasmids, which do not inflict costs on their hosts, there would "$!
be no additional fitness cost from a gene being carried on the plasmid compared to the "%!
chromosome. Small elements, such as very small plasmids, may not exert a "&!
discernible fitness effects on their host if they do not have large copy-numbers. If "'!
there are no costs to a mobile element, it will be able to spread in a population through "(!
either horizontal gene transfer or, in the case of small populations, genetic drift. In ")!
fact, reduced costs of plasmid carriage may have been the driving force for the "*!
creation of secondary chromosomes in many bacteria. Such chromosomes often have #+!
plasmid-like features and are ubiquitous in some bacterial clades (Egan and Waldor, #"!
2003; Slater et al, 2009). Accordingly, recent work shows that the very largest ##!
plasmids have lost mobility and acquired essential genes (Smillie, Rocha, de la Cruz #$!
in preparation). #%!
#&!
13!
"!
Mobile genetic elements as mutualists #!
MGEs have been associated with the spread of a wide range of adaptive traits that $!
enhance the fitness of their hosts. For example, plasmids commonly carry resistance %!
genes that enable bacteria to grow in the presence of antibiotics or heavy metals, &!
which tend to have spatially or temporarily variable distributions (Eberhard, 1990). '!
Plasmids also have important roles in the establishment of antagonisms or mutualisms (!
between prokaryotes and eukaryotes, such as virulence traits ,EF/B261G12!"#!$%8!)!
#+++=, nitrogen fixation by genes coded in rhizobia symbiotic plasmids (Nuti et al, *!
1979) or amino acid production in Buchnera (Gil et al, 2006; Nuti et al, 1979). "+!
Phages have been shown to be important in bacterial virulence, as in the case of ""!
cholera (Faruque et al, 2005), or in mutualisms, as in the case of an aphid-bacterial "#!
association (Oliver et al, 2009). "$!
"%!
Genes carried on plasmids may be readily transferred to the bacterial chromosome. "&!
Theoretical models suggest that constant selection pressure should favour transfer of "'!
beneficial plasmid genes to the chromosome to avoid the costs of plasmid carriage "(!
(Bergstrom et al, 2000; Eberhard, 1990). This raises the question as to why beneficial ")!
genes remain on mobile elements, rather than integrating themselves onto the "*!
chromosome. If selection pressures vary over time or space, genes which are #+!
beneficial in some environments, but not others, will be able persist on MGEs. In a #"!
spatially structured environment, mobile elements help to facilitate the transfer of ##!
beneficial traits which have previously evolved in local populations, to other sub-#$!
populations (Bergstrom et al, 2000). In addition, it has been argued that locally-#%!
adapted genes which are carried on plasmids allow the persistence of these traits in #&!
14!
the face of selective sweeps (Bergstrom et al, 2000; Turner et al, 1998), allowing for "!
the generation of more favourable gene combinations to arise on plasmids. #!
$!
Parasites and symbionts can often be seen as lying on a continuum between harming %!
(in the case of parasites) and helping (in the case of mutualists) their partners. Mobile &!
elements that code for beneficial genes can be seen as mutualists, and should face the '!
same evolutionary dilemmas (Ferdy and Godelle, 2005; Foster and Wenseleers, 2006; (!
Sachs and Simms, 2007). When should a mobile element provide a benefit and when )!
should it harm its host, and how will these evolutionary decisions in turn shape *!
mobility? The answer to these questions are likely to depend on the details of the "+!
tradeoffs between horizontal transmission, vertical transmission and virulence (Ferdy ""!
and Godelle, 2005; Haft et al, 2009). The more an element transmits vertically, the "#!
more it will depend on the reproductive value of the host, and selection will therefore "$!
favour genes which are beneficial to the host, while higher horizontal gene transfer "%!
rates will act to the detriment of the host (subject to tradeoffs). Over long time-scales, "&!
elements which harm the host are likely to be lost, and those which are beneficial to "'!
the host will likely be integrated into the host genome. "(!
")!
Despite the potential benefits of some plasmids to the organisms involved, sharing "*!
beneficial DNA with other members of the population can be seen as a social #+!
dilemma. The reason for the dilemma is that an organism which does not suppress the #"!
transmission of a plasmid will be benefiting its potential competitors. If transmitting a ##!
plasmid is costly, then this may be seen as an altruistic act. In this way, horizontal #$!
gene transfer has been suggested to be analogous to teaching in the spread of culture #%!
in humans (Riboli-Sasco et al, 2008). Behaviours learned from teaching are similar to #&!
15!
plasmids in the sense that transfer is costly, they frequently confer a benefit on the "!
recipient, and the recipient can further transmit the information to other individuals #!
(Riboli-Sasco et al, 2008). $!
%!
Mobile genetic elements as drivers of bacterial sociality &!
Our figure 2 follows from W.D. Hamilton’s classical categorization of social '!
behaviours (Hamilton, 1964; West et al, 2007b; West et al, 2007c), where social (!
behaviours are categorized by their net lifetime direct effect on the fitness of a focal )!
bacterium (the actor) and on neighbouring bacteria (the recipients). The resulting four *!
behaviours are “selfishness” (+/-) which confers a benefit on the actor, whilst "+!
inflicting a cost on the social partner, “altruism” (-/+), “spite” (-/-) and “mutually ""!
beneficial behaviours” (+/+ – Hamilton, 1964; West et al, 2007c). Over the last "#!
decade, a growing body of work has been devoted to microbial sociality, and all four "$!
of these social interactions can be observed in microbial populations (Crespi, 2001; "%!
Griffin et al, 2004; Rainey and Rainey, 2003; West et al, 2007a; West et al, 2006; "&!
Xavier and Foster, 2007). We assert that, when looking at social behaviours, all of "'!
these four behaviours can be seen to be encoded by horizontally transferred genes. We "(!
further argue that mobile elements, due to their effect on the local genetic structure, ")!
lend themselves to promoting cooperative social traits. "*!
#+!
Cooperation: Altruism and mutually beneficial traits #"!
Bacteria are remarkably cooperative, producing a diversity of shared, secreted ##!
products (public goods) that can enhance growth in a diversity of challenging #$!
environments (West et al, 2007a; West et al, 2006). However, like any social #%!
organism, cooperative bacteria are always vulnerable to exploitation by non-#&!
16!
producing cheats. In the absence of any kin or spatial structure, any individuals in a "!
population which do not produce a public good will have an advantage over #!
individuals which do, a situation referred to as the tragedy of the commons (Hardin, $!
1968; Rankin et al, 2007). As non-producing individuals have a fitness advantage %!
over those that produce public goods, non-producing individuals will invade over &!
time. '!
(!
Many traits which have been shown to be involved in bacterial cooperation and )!
virulence are coded by mobile elements, such as traits which are capable of breaking *!
down toxins in the local environment, (Ellis et al, 2007; Knothe et al, 1983; Lee et al, "+!
2006; Philppon et al, 2002) and secreted toxins (Ahmer et al, 1999; O’Brien et al, ""!
1984; Waldor and Mekalanos, 1996). One theoretical study (Smith, 2001) suggested "#!
that if cooperation is coded on plasmids, these plasmids will be able to reinstate "$!
cooperation by forcing non-producing strains (via infection) to start secreting the "%!
extracellular public good (Smith, 2001). However, once such a cooperative plasmid "&!
invaded a population of non-cooperating individuals, the population would still be "'!
prone to invasion from an incompatible (i.e. rival) plasmid, which did not code for "(!
cooperation. Thus, over time it is likely that cooperation would break down, as the ")!
social dilemma repeats itself at the level of plasmids (McGinty, Rankin & Brown, in "*!
preparation; Nogueira et al. 2009). #+!
#"!
Figure 1 illustrates the nested population structure of bacterial plasmids, where ##!
plasmids are nested within bacteria, and these bacteria, in turn, are nested within #$!
patches (e.g. hosts or other resource pools). As plasmids can spread horizontally #%!
within a local population, they have the potential to change the genetic structuring #&!
17!
among individuals in local populations. Relatedness is an influential measure of "!
population genetic structure, used in particular to decipher the direction of selection #!
on social traits (Griffin et al, 2004; Hamilton, 1964; Rousset, 2004; West et al, 2006), $!
with higher relatedness between individuals tending to favour the evolution of %!
cooperative traits (Hamilton, 1964). Relatedness should properly be measured for a &!
focal locus directly. Box 2 describes a commonly used method to calculate '!
relatedness, and demonstrates that relatedness at mobile loci can increase as a result of (!
horizontal gene transfer. While spatial structure, leading to increased genetic )!
relatedness among bacterial hosts, can lead to cooperative traits on the host *!
chromosome, they will be more strongly favoured on plasmids, as the nature of "+!
horizontal gene transfer can act to increase genetic relatedness even more at mobile ""!
loci (figure 3). As such, it is expected that many genes involved in the production of "#!
cooperative traits, or public goods, will be carried by MGEs, in particular by the most "$!
transmissible conjugative plasmids. "%!
"&!
Genes coding for proteins which are secreted outside of the cell can be referred to as "'!
being part of the secretome (Ahmer et al, 1999; Nogueira et al, 2009; Sanchez et al, "(!
2008). As proteins are costly to produce, secreting them outside of a cell has a clear ")!
cost to the focal individual. Analysing the genomes of 20 Escherichia and Shigella "*!
lineages, and investigating where proteins were likely to be expressed within a cell, #+!
Nogueira et al. (2009) were able to identify which genes coded for secreted proteins. #"!
The genomes were further analysed, looking for evidence of transfer hotspots (areas ##!
in the genome more likely to be transmitted horizontally) and identifying whether the #$!
genes coding for secreted proteins were more likely to be found on transfer hotspots #%!
or plasmids. Among genes for which localization could be predicted, it was found that #&!
18!
8% of chromosomal hot-spot genes coded for extracellular proteins, a higher "!
percentage than for cold-spots (6%), the least mobile gene class. Furthermore, 15% of #!
plasmid genes – the most mobile class of genes in the analysis – coded for $!
extracellular proteins (Nogueira et al, 2009). This suggests that horizontal gene %!
transfer is an important factor in the evolution of social behaviours in microbes &!
(Nogueira et al, 2009). '!
(!
Another example of plasmids influencing the social environment comes from the )!
tumour inducing (Ti) plasmid in the bacterium Agrobacterium tumefaciens, a plant *!
pathogen. For the bacterium to be virulent, it must carry a copy of the Ti plasmid, "+!
which is then inserted into the cells of plants. Once inside, the plasmid induces cell ""!
division, which creates a gall (Zupan et al, 2000). The gall releases opines which can "#!
then be used as an important source of nitrogen and energy for the bacterium (White "$!
and Winans, 2007; Zupan et al, 2000). An interesting twist on to the tale is that only "%!
bacteria which are infected with the plasmid can utilize the opines, meaning that "&!
cooperation by plasmids only favours other individuals which contain the plasmid, "'!
making the trait a so-called “greenbeard” (Gardner and West, In Press). "(!
")!
Spiteful traits "*!
The flipside of altruism is spite, where an individual pays a cost to inflict another cost #+!
on another individual (Gardner and West, 2004a; Gardner and West, 2004b; Gardner #"!
and West, 2006; Hamilton, 1970). This is particularly common in the case of bacteria, ##!
where individuals may produce bacteriocins, or other toxins, to kill their conspecifics #$!
(Ackermann et al, 2008; Dykes and Hastings, 1999; Gardner et al, 2004; Riley and #%!
Wertz, 2002). Spite has long been seen as a puzzling evolutionary phenomenon, as a #&!
19!
given actor pays a net cost -c to inflict a net cost -b on another member of the "!
population (Foster et al, 2001). Spite occurs in a number of systems (Foster et al, #!
2000; Gardner et al, 2007; Gardner et al, 2004; Keller and Ross, 1998) but has most $!
predominantly been invoked with respect to anti-competitor behaviours in bacteria %!
(Brown et al, 2009a; Dionisio, 2007; Gardner et al, 2004; West et al, 2007a; West et &!
al, 2006). '!
(!
Spiteful extracellular products, which are costly for the producer and cause harm to )!
other members of the population, are frequently found on mobile elements, *!
particularly plasmids and phages (e.g. Brown et al, 2006). In the case of bacteriocins, "+!
plasmids can carry both toxins and the corresponding genes for resistance to the toxin ""!
(Riley and Wertz, 2002). Spite can evolve if the individual acting spitefully and the "#!
individual on the receiving end of the spiteful behaviour are negatively related (i.e. if "$!
r<0). In other words, as the average relatedness in a population must be r=0, the "%!
respective individuals must be less related than average, so damaging these "&!
individuals will increase the proportion of individuals carrying the focal actors genes. "'!
Usually this occurs under small population sizes (Gardner and West, 2004b) or if "(!
there is a way of recognizing unrelated individuals (Brown and Buckling, 2008; ")!
Keller and Ross, 1998). If bacteriocins are carried on plasmids, then this becomes "*!
easier: if a cell coding for bacteriocins lyses, it kills all members of the #+!
neighbourhood which do not bear the plasmid gene, and thus favours individuals #"!
which carry the plasmid (as they also carry resistance to the bacteriocin). As such, one ##!
may regard such genes as being ‘greenbeards’, as bacteriocinogenic individuals #$!
(carriers of the toxin-immunity gene complex) preferentially help other individuals #%!
carrying the exact same complex to survive (Gardner and West, In Press). Thus, the #&!
20!
plasmid-carried bacteriocins may be seen as analogous to the plasmid addiction "!
complexes discussed previously, as they are mechanisms which favour the genes #!
being on plasmids but, unlike addiction complexes (which work within the cell to kill $!
individuals which do not carry the plasmid), they work by killing neighbouring cells. %!
&!
We should expect bacteriocins to be carried by plasmids at intermediate levels of '!
horizontal gene transfer, or higher rates of segregation: if transfer is too high, and all (!
individuals in a local neighbourhood carry a plasmid, then there are no non-carriers to )!
kill. If in contrast plasmids are too rare, then there would be insufficient toxin and *!
subsequent killing to compensate for the fixed costs of toxin production (Chao and "+!
Levin, 1981). The targets of bacteriocins are not always the same bacterial strain ""!
,5.G/.71G!"#!$%8!#++(=. This presents an additional function for bacteriocin-carrying "#!
plasmids: if they are less efficient at infecting other strains, any plasmids which "$!
produce toxins to kill those strains will have an advantage from reducing competition "%!
with the strain it can more efficiently infect. However, as for other social traits, this "&!
will very much depend on the genetic structure of the local environment. "'!
"(!
Conclusions ")!
Our review has highlighted the fact that a diversity of traits can be coded in mobile "*!
elements. Among this diversity, we argue that there are systematic biases towards the #+!
carriage of certain traits. It is clear that mobile elements frequently evolve to confer #"!
benefits on their bacterial hosts (e.g. Bergstrom et al, 2000; Faruque et al, 2005; ##!
Oliver et al, 2009; Philppon et al, 2002; Slater et al, 2008). It is also particularly #$!
striking that a wide range of secreted products can be found to be coded in the mobile #%!
part of the genome (Nogueira et al, 2009), highlighting the importance of mobile #&!
21!
elements as agents of cooperation and conflict in microbial populations, and as model "!
systems in which to study questions related to social evolution. With the growing #!
number of bacterial and plasmid genomes becoming fully available, testing questions $!
regarding social evolution with large genomic data sets will become increasingly %!
feasible. &!
'!
Many genes which are carried by MGEs are involved in both the evolution of (!
antibiotic resistance and bacterial virulence. Therefore, understanding the conditions )!
under which these genes are carried on mobile genetic elements is essential if we are *!
to get a better understanding of how pathogenic bacteria evolve, and how potentially "+!
we can control them (Martinez, 2009; Martínez, 2008).Understanding the selective ""!
pressures which favour genes to be carried on plasmids could be useful to evaluate "#!
whether controlling plasmid transmission is an effective long-term mechanism to "$!
counter the evolution of antibiotic resistance (Amabile-Cuevas and Heinemann, "%!
2004). The emerging importance of MGEs for the expression of bacterial cooperative "&!
phenotypes (Nogueira et al. 2009) opens up a further arena of potential biomedical "'!
application. Understanding the link between bacterial cooperation and virulence "(!
,-.226G;3!"#!$%8!#++'= can in principle contribute to new medical intervention ")!
strategies, such as the use of genetically engineered ‘Trojan cheat’ bacteria to invade "*!
and undermine established cooperative foci of infection ,E2;H3!"#!$%8!#++*I=. The #+!
realization that cooperative traits are often plasmid-encoded opens the analogous #"!
possibility of introducing ‘Trojan plasmids’ to undermine the cooperative virulence of ##!
the focal population and/or drive medically useful alleles into the focal population that #$!
favour secondary mechanisms of control. #%!
22!
"!
Acknowledgements #!
We are grateful to the Swiss National Science Foundation (grants 31003A-125457 $!
and PZ00P3-121800 to DJR)!and the University of Zürich Forschungskredit (DJR), %!
the CNRS and the Institut Pasteur (EPCR), and the Wellcome Trust (grant &!
082273/Z/07/Z to SPB) for funding. We also thank Sorcha McGinty for comments on '!
the manuscript.(!
23!
Box 1. Modeling plasmid persistence "!
Susceptible-infected (SI) models have been useful in studying infectious diseases #!
(Anderson and May, 1979; May and Anderson, 1979) and additionally have been $!
applied to the spread of cultural traits (Riboli-Sasco et al, 2008). Such models also %!
provide powerful tools for studying the ecology and evolution of plasmids (Bergstrom &!
et al, 2000; Lili et al, 2007; Smith, 2001). By ignoring copy number within cells, and '!
simply assuming that a given cell will either be susceptible to plasmid infection or (!
infected by a plasmid, we can model the dynamics as follows: )!
dF/dt=F(a(1–N/K) + bP/N – !P – !Q) + sP +sQ (1b) *!
dP/dt=P(a(1–N/K) + bP/N + !F – u – c – s) (1a) "+!
Here F refers to the density of cells without a plasmid (“suscepitble” cells) and P ""!
refers to the density of cells which contain a plasmid (“infected” cells). Here we "#!
assume that the plasmid represented by P code for a product, at a cost c. We can "$!
assume that there are other plasmids in the population, whose density is given by Q "%!
which are incompatible with P plasmids, and do not pay the cost c. The rate of their "&!
growth is given by "'!
dQ/dt=Q(a(1–N/K) + bP/N + !F – u – s) (2) "(!
The overall density of cells is N=F+P+Q, the growth rate is given by a, the rate of ")!
horizontal gene transfer is given by !, x is the cost of conjugation and s is the rate of "*!
plasmid loss (segregation). The parameter c describes the costs (if c>0, or benefit if #+!
c<0) from any genes carried on the plasmid, and b describes the effect of secreted #"!
products produced by a given plasmid on the population as a whole. The condition for ##!
plasmids to invade from rare is given by (dP/dt)/P>0. Evaluated for the case when #$!
plasmid-free cells are at equilibrium density and plasmids are rare (i.e. P!0 and #%!
24!
N*=F*=K), the condition for plasmids to invade is !K>c+u+s, and is thus "!
independent of the population growth rate, or the effect of secreted proteins b which #!
are produced (as their effect in a large population is negligible population size is $!
assumed to be large). Thus, if secreted proteins come at a cost c, the cost must be %!
outweighed by horizontal gene transfer or spatial structure. &!
'!
(!
25!
Box 2. Relatedness and social evolution "!
Kin selection theory predicts that a trait which confers a net cost c on a focal #!
individual, and confers a benefit b on a given social partner will be able to evolve if $!
rb–c>0 (3) %!
where r is the genetic relatedness between individuals. This is known as Hamilton’s &!
rule (Hamilton, 1964). Altruism requires that r>0, b>0 and c>0). Interestingly, this '!
inequality can additionally explain spiteful interactions between individuals, where a (!
gene confers a cost c on its bearer, and an additional cost b (where b<0), as long as the )!
relatedness between interacting individuals is r<0. *!
"+!
Horizontal gene transfer can affect relatedness between individuals (Nogueira et al, ""!
2009). As relatedness is always measured at the locus in question, infecting "#!
individuals in a local patch with an MGE will increase the relatedness of "$!
neighbouring bacterial cells, measured at MGE loci. "%!
"&!
In the absence of horizontal gene transfer, relatedness can be calculated from the size "'!
N of a local patch, and the probability m that a given individual will migrate in a given "(!
time step t (Rousset, 2004). Defining relatedness as the probability that any two ")!
randomly picked individuals are identical at the locus of interest, relatedness can be "*!
given as #+!
R t +1( ) = 1!m( )2 1N
+ N !1N
R t( )" # $
% & ' (4)
#"!
where (1-m)2 is the probability that two random individuals will remain in a patch ##!
during a given time-interval and 1/N is the probability that the two non-migrant #$!
individuals stem from the same parent in the previous time-step (Rousset, 2004). #%!
26!
"!
At equilibrium, the relatedness between individuals in a patch is #!
R* =1!m( )2
N ! 1!m( )2 N !1( ). (5) $!
Drawing form models of cultural evolution (Lehmann et al, 2008), we now extend %!
this recursion to allow for horizontal gene transfer. We can expect that horizontal &!
gene transfer (e.g via plasmid conjugation) will affect relatedness within a patch, '!
either through the plasmid infecting other individuals (and thus increasing (!
relatedness) or through plasmid loss (due to segregation). The probability that two )!
individuals carrying distinct alleles become identical in the next time-step (through *!
horizontal gene transfer) is pG and the probabiity that two individuals which carry "+!
identical alleles at the focal locus remain identical (through neither losing the gene) in ""!
the next time-step is pK. "#!
R t +1( ) = 1!m( )2 pKN
+ N !1N
pKR t( ) + pG 1! R t( )( )( )" # $
% & ' + pG 1! 1!m( )2( ) (6) "$!
The overall rate of horizontal gene transfer will depend on the frequency R(t) of cells "%!
in the patch which are infected with the focal mobile element, as well as the "&!
probability ! that the mobile element will be transmitted given that two cells (one "'!
infected and one uninfected with the element) meet. We can therefore write pG=!R(t). "(!
We also assume that a given element is lost with probability s, and the probability that ")!
a given cell which had an element at time t will still have the element at time t is "*!
therefore pK=1-s. When equation (6) is at the equilibrium R*, where R(t+1)=R(t), #+!
segregation reduces relatedness, while transmission increases relatedness: if s=0 and #"!
m=0 (i.e. there is no migration between patches, and plasmids are never lost) then ##!
relatedness converges to 1, as all cells eventually become infected with the MGE. If #$!
27!
we assume very large local population sizes (and thus n!") we can derive an "!
analytical solution for (6) which is: #!
R* =! " s " 1" s( )m 2 "m( )
! 1"m( )2 (7)
$!
This shows that increasing ! always increases R: horizontal transfer (infection) of %!
MGEs increases relatedness within patches, while segregation reduces relatedness &!
(figure 3) when the focal gene is carried on a mobile genetic element. As increased '!
local relatedness favours cooperation (Hamilton, 1964; Lehmann and Keller, 2006; (!
Sachs et al, 2004), we conclude that horizontally-transferred genes will be more likely )!
to code for cooperative traits than those that are less infectiously mobile. *!
"+!
""!
28!
References "!
J/B@<.3!K!,"*(&=>!K.@63L!.LL21L.@1G!63!&'()"*+()+$!(,%+!/;3MFL.@6;3>!-,.*/$%!,0!#!
1$(#"*+,%,23!!"#,#=$!&+&N&"&>!$!
!%!
J/012<.33!K8!O@1/B12!E8!P2114!:Q8!O;3LB1@!C8!-.24@!RNS8!S;1I176!K!,#++)=>!&!
O17TN41G@2F/@6U1!/;;V12.@6;3!<146.@14!IW!VB13;@WV6/!3;6G1>!4$#.*"!%&%$!*)(X'!
**+>!(!
!)!
JB<12!EKK8!Y2.3!K8!-12;3!P!,"***=>!YB1!U62F713/1!V7.G<64!;T!O.7<;3177.!*!
@WVB6<F26F<[email protected]<6GG6I71>!-,.*/$%!,0!5$(#"*+,%,23!!'!$!"$'%N"$')>!"+!
!""!
J76D;3!O8!-F2T;24!J8!K641;!:8!U.3!E..713!K!,#++*=>!Z62F713/1!1U;7F@6;3!.34!@B1!"#!
@2.41N;TT!BWV;@B1G6G[!B6G@;2W8!/F2213@!G@.@1!;T!.TT.62G!.34!@B1!TF@F21>!>!-,.*/$%!,0!"$!
&6,%.#+,/$*3!5+,%,23!""$!#%&N#&*>!"%!
!"&!
J<.I671N5F1U.G!5P8!-1631<.33!\J!,#++%=>!OB;;@63L!@B1!<1GG13L12!;T!.3@6I6;@6/!"'!
[email protected]/1[!V7.G<64!176<63.@6;3!.G!.!/;F3@12N1U;7F@6;3.2W!@./@6/>!7*.2!7+'(,6"*3!"(!
8,9$3!(,""=$!%'&N%'(>!")!
!"*!
J3412G;3!]K8!K.W!]K!,"*(*=>!C;VF7.@6;3!I6;7;LW!;T!63T1/@6;FG!46G1.G1G[!C.2@!^>!#+!
4$#.*"!"'),&(#"=$!$'"N$'(>!#"!
!##!
J3412G;3!]K8!K.W!]K!,"**#=>!:/0"(#+,.'!7+'"$'"'!,0!;.<$/'=!73/$<+('!$/9!#$!
>,/#*,%!A_T;24!`36U12G6@W!C21GG[!A_T;24>!#%!
!#&!
29!
J2/FG!Z?8!].631W!CE8!YF2312!O\!,#++&=>!YB1!C^:N4;<.63!@;_63N.3@6@;_63!.22.W!63!"!
<W/;I./@126.>!8*"/9'!+/!?+(*,1+,%,23!!#,)=$!$'+N$'&>!#!
!$!
E.B7!K^8!-.3G13!?-8!?6/B@!Y]8!O;213G13!O\!,#++(.=>!5;3MFL.@6U1!Y2.3GT12!%!
P./676@.@[email protected][email protected]/1!;T!^3/CN"!C7.G<64!Va\a&!63!QG/B126/B6.!/;76!5177G!&!
5;7;36D63L!@B1!b.G@2;63@[email protected]!Y2./@!;T!@B1!b12<T211!].@>!@AA%+"9!$/9!'!
&/6+*,/<"/#$%!?+(*,1+,%,23!*#,"=$!$%"N$%$>!(!
!)!
E.B7!K^8!?.2G!-1G@IM12L!-8!Oc213!\O!,#++(I=>!^<V./@!;T!/;[email protected]!;3!@B1!*!
[email protected]@W!;T!^3/CN"!V7.G<64!Va\a&!63!I./@126.7!V;VF7.@6;3G>!B&?C!?+(*,1+,%,23!"+!
D"##"*'!"++,#=$!#&+N#&'>!""!
!"#!
E.22.3L;F!]8!P21<.F_!58!S1U1.F!-8!]6/B.24G!K8!E;W.U.7!C8!K;631.F!O!"#!$%!"$!
,#++(=>!5]^OC]!V2;U641G!./[email protected]/1!.L.63G@!U62FG1G!63!V2;0.2W;@1G>!"%!
C(+"/("!#!&$!"(+*N"("#>!"&!
!"'!
E12LG@2;<!5Y8!?6VG6@/B!K8!?1U63!E]!,#+++=>!:[email protected]!O171/@6;38!^3T1/@6;FG!Y2.3GT12!"(!
.34!@B1!Q_6G@13/1!5;346@6;3G!T;2!E./@126.7!C7.G<64G!E"/"#+('!!&&$!"&+&N"&"*>!")!
!"*!
E;2413G@163!O]8!]1D360;TT!RO!,#++&=>!K;I671!S:J!63!;I76L.@1!63@2./177F7.2!#+!
I./@126.>!4$#.*"!F"6+"G'!?+(*,1+,%,23!#,*=$!'))N'**>!#"!
!##!
E;F<.!\Q8!?13G06!]Q!,"*))=>!QU;7F@6;3!;T!.!E./@126.!C7.G<64!JGG;/6.@6;3>!4$#.*"!#$!
##&,'"))=$!$&"N$&#>!#%!
!#&!
30!
E2;H3!OC8!EF/0763L!J!,#++)=>!J!G;/6.7!76T1!T;2!46G/12363L!<6/2;I1G>!>"%%!!#&,-+))."!
+)%>!#!
!$!
E2;H3!OC8!^3L76G!]P8!Y.4416!P!,#++*.=>!QU;7F@6;3.2W!1/;7;LW!;T!<6/2;I6.7!H.2G[!%!
H6@B63NB;G@!/;<V1@6@6;3!.34!,63/[email protected]=!U62F713/1>!&6,%.#+,/$*3!@AA%+($#+,/'!&!
","=$!$#N$*>!'!
!(!
E2;H3!OC8!?1!5B.@!?8!S1!C.1V1!K8!Y.4416!P!,#++'=>!Q/;7;LW!;T!<6/2;I6.7!)!
63U.G6;3G[!J<V76T6/.@6;3!.77;HG!U62FG!/.22612G!@;!63U.41!<;21!2.V647W!HB13!2.21>!*!
>.**"/#!5+,%,23!!+,#+=$!#+%)N#+&#>!"+!
!""!
E2;H3!OC8!R1G@!OJ8!S6LL71!OC8!b26TT63!JO!,#++*I=>!O;/6.7!1U;7F@6;3!63!<6/2;N"#!
;2L.36G<G!.34!.!Y2;M.3!B;2G1!.VV2;./B!@;!<146/.7!63@12U13@6;3!G@2.@1L61G>!"$!
H)+%,',A)+($%!8*$/'$(#+,/'!,0!#)"!F,3$%!C,(+"#3!5=!5+,%,2+($%!C(+"/("'!#+%,"&$$=$!"%!
$"&(N$"')>!"&!
!"'!
EF/B261G12!58!b7.G12!C8!]FG36;0!58!:14M.26!-8!Se-.F@1U6771!-8!aF3G@!P!"#!$%!,#+++=>!"(!
YB1!U62F713/1!V7.G<64!VR]"++!.34!@B1!21V12@;621!;T!V2;@163G!G1/21@14!IW!@B1!")!
@WV1!^^^!G1/21@6;3!.VV.2.@FG!;T!GB6L177.!T71_3126>!?,%"(.%$*!<+(*,1+,%,23-#'$!('+N"*!
((">!#+!
!#"!
EFGB<.3!P!,#++#=>!D$#"*$%!74@!8*$/'0"*=!?"()$/+'<'!$/9!>,/'"I."("'>!5;74!##!
OV263L!-.2I;F2!`36U12G6@W!C21GG[!5;74!OV263L!-.2I;F28!:f>!#$!
!#%!
31!
EF@G!?8!?.B!\8!S.;NYB6!KN-8!RW3G!?8!?;26G!]!,#++&=>!Y;_63N.3@6@;_63!<;4F71G!.G!"!
I./@126.7!<[email protected];76/!G@21GG!<.3.L12G>!8*"/9'!+/!5+,()"<+($%!C(+"/("'!#),"#=$!'(#N#!
'(*>!$!
!%!
5.G/.71G!Q8!EF/B.3.3!Oa8!SF/B1!S8!a71.3@B;FG!58!?7;FI1G!]8!C;G@71!a!"#!$%!,#++(=>!!&!
!'!
5;76/63!I6;7;LW>!?+(*,1+,%3!$/9!?,%"(.%$*!5+,%,23!F"6+"G!*!$!"&)N##*>!(!
!)!
5.G/.71G!Q8!5B26G@61!C\!,#++$=>!YB1!U12G.@671!I./@126.7!@WV1!^Z!G1/21@6;3!GWG@1<G>!*!
4$#.*"!F"6+"G'!?+(*,1+,%,23!!$!"$(N"%*>!"+!
!""!
5B.;!?8!?1U63!E]!,"*)"=>!O@2F/@F214!B.I6@.@G!.34!@B1!1U;7F@6;3!;T!.3@6/;<V1@6@;2!"#!
@;_63G!63!I./@126.>!H*,(""9+/2'!,0!#)"!4$#+,/$%!@($9"<3!,0!C(+"/("'!*'$!'$#%N"$!
'$#)>!"%!
!"&!
5B13!^8!SFI3.F!S!,#++%=>!S:[email protected]!4F263L!I./@[email protected];2<.@6;3>!4$#.*"!"'!
F"6+"G'!?+(*,1+,%,23!"$!#%"N#%*>!"(!
!")!
5;;V12!YP8!-1631<.33!\J!,#+++=>!C;G@G1L21L.@6;3.7!067763L!4;1G!3;@!63/21.G1!"*!
V7.G<[email protected]@W!IF@!./@G!@;!<146.@1!@B1!1_/7FG6;3!;T!/;<V1@6@63L!V7.G<64G>!#+!
H*,(""9+/2'!,0!#)"!4$#+,/$%!@($9"<3!,0!C(+"/("'!,0!#)"!J/+#"9!C#$#"'!,0!@<"*+($!#"!
(*,#$=$!"#'%$N"#'%)>!##!
!#$!
5;;V12!YP8!-1631<.33!\J!,#++&=>!O171/@6;3!T;2!V7.G<64!V;G@NG1L21L.@6;3.7!#%!
067763L!41V134G!;3!<F7@6V71!63T1/@6;3[!1U6413/1!T;2!@B1!G171/@6;3!;T!<;21!U62F713@!#&!
32!
V.2.G6@1G!@B2;FLB!V.2.G6@1N71U17!/;<V1@6@6;3>!H*,(""9+/2'!,0!#)"!F,3$%!C,(+"#3!,0!"!
D,/9,/!C"*+"'!5K5+,%,2+($%!C(+"/("'!"*"$!%+$N%"+>!#!
!$!
521GV6!E\!,#++"=>!YB1!1U;7F@6;3!;T!G;/6.7!I1B.U6;2!63!<6/2;;2L.36G<G>!8*"/9'!+/!%!
&(,%,23!L!&6,%.#+,/!!+,%=$!"()N")$>!&!
!'!
41!7.!52FD!P8!P2;G@!?O8!K1W12!]\8!g1/B312!Q!,#++*=>!5;3MFL.@6U1!S:[email protected];76G<!(!
63!b2.<N31L.@6U1!E./@126.>!B&?C!?+(*,1+,%,%,23!F"6+"G'!#%$!")N%+>!)!
!*!
S17GF/!P8!E2630<.33!-8!CB676VV1!-!,#++&=>!CBW7;L13;<6/G!.34!@B1!"+!
21/;3G@2F/@6;3!;T!@B1!@211!;T!76T1>!4$#!F"6!E"/"#!+,&=$!$'"N$(&>!""!
!"#!
S6.D!]6//6!\58!-123.341D!KQ!,#+++=>!C7.G<64!1TT1/@G!;3!QG/B126/B6.!/;76!"$!
<[email protected];76G<>!>*+#+($%!F"6+"G'!+/!5+,#"()/,%,23!"),#=$!(*N"+)>!"%!
!"&!
S6;36G6;!P!,#++(=>!O17T6GB!.34!GV6@1TF7!I1B.U6;F2!@B2;FLB!V.2.G6@1G!.34!"'!
V.@B;L13G>!&6,%.#+,/$*3!&(,%,23!F"'"$*()!($!""**N"#"+>!"(!
!")!
S6;36G6;!P8!K.@6/!^8!].4<.3!K8!];426LF1G!A]8!Y.4416!P!,#++#=>!C7.G<64G!GV21.4!"*!
U12W!T.G@!63!B1@12;L131;FG!I./@126.7!/;<<F36@61G>!E"/"#+('!!+",%=$!"&#&N"&$#>!#+!
!#"!
SW01G!bJ8!-.G@63LG!\R!,"***=>!O171/@6;3!.34!T6@31GG!63!I./@126;/63NV2;4F/63L!##!
I./@126.>!H*,(""9+/2'!,0!#)"!F,3$%!C,(+"#3!,0!D,/9,/!C"*+"'!5K5+,%,2+($%!C(+"/("'!#$!
"+%,"$)#=$!')$N')(>!#%!
!#&!
33!
QI12B.24!Rb!,"**+=>!QU;7F@6;3!63!I./@126.7!V7.G<64G!.34!71U1G!;T!G171/@6;3>!8)"!"!
M.$*#"*%3!F"6+"G!,0!5+,%,23!+&,"=$!$N##>!#!
!$!
QL.3!QO8!R.74;2!Ka!,#++$=>!S6G@63/@!21V76/.@6;3!21dF621<13@G!T;2!@B1!@H;!N+1*+,!%!
(),%"*$"!/B2;<;G;<1G>!>"%%!!!%,%=$!&#"N&$+>!&!
!'!
Q776G!]\8!?6771W!Ja8!?./1W!O\8!KF22177!S8!b;4T2.W!-5\!,#++(=>!P21dF13/WN41V13413@!(!
[email protected]!;T!V7.G<64!/.226.L1!IW!CG1F4;<;3.G!63!B;<;L131;[email protected]!)!
G@2F/@F214!13U62;3<13@G>!:'<"!-,.*/$%!!,"=$!*#N*&>!*!
!"+!
P.2FdF1!OK8!:.G12!^E8!^G7.<!K\8!P.2FdF1!JOb8!bB;GB!J:8!:.62!bE!"#!$%!,#++&=>!""!
O1.G;3.7!1V641<6/G!;T!/B;712.!63U12G17W!/;2217.@1!H6@B!@B1!V21U.713/1!;T!"#!
13U62;3<[email protected]!/B;712.!VB.L1G>!H*,(""9+/2'!,0!#)"!4$#+,/$%!@($9"<3!,0!C(+"/("'!,0!"$!
#)"!J/+#"9!C#$#"'!,0!@<"*+($!!)",&=$!"(+#N"(+(>!"%!
!"&!
P124W!\E8!b;41771!E!,#++&=>!S6U12G6T6/.@6;3!;[email protected]<6GG6;3!<;41G!.34!@B1!"'!
1U;7F@6;3!;T!<F@[email protected]<>!@<"*+($/!4$#.*$%+'#!!++,&=$!'"$N$#(>!"(!
!")!
P;G@12!a]8!].@3610G!P?R8!R13G17112G!Y!,#+++=>!OV6@1!63!G;/6.7!63G1/@G>!8*"/9'!+/!"*!
&(,%,23!L!&6,%.#+,/!!&,""=$!%'*N%(+>!#+!
!#"!
P;G@12!a]8!R13G17112G!Y!,#++'=>!J!L1312.7!<;417!T;2!@B1!1U;7F@6;3!;T!##!
<[email protected]<G>!-,.*/$%!,0!&6,%.#+,/$*3!5+,%,23!!($!"#)$N"#*$>!#$!
!#%!
34!
P;G@12!a]8!R13G17112G!Y8!].@3610G!P?R!,#++"=>!OV6@1[!-.<67@;3eG!F3V2;U13!"!
@B1;2W>!@//$%"'!O,,%,2+(+!B"//+(+!#',$N%=$!##*N#$)>!#!
!$!
P;_!]Q8!gB;3L!h8!a2;31!OK8!Y;V!QK!,#++)=>[email protected]!G@2F/@F21!.34!3F@2613@G!%!
V2;<;@1!63U.G6;3!;T!^3/CN"!V7.G<64G!63!I./@126.7!V;VF7.@6;3G>!8)"!:C?&!-,.*/$%!&!
"$!"+#%N"+$*>!'!
!(!
P2;G@!?O8!?1V7.1!]8!OF<<12G!JA8!Y;FGG.63@!J!,#++&=>!K;I671!L131@6/!171<13@G[!)!
@B1!.L13@G!;T!;V13!G;F2/1!1U;7F@6;3>!4$#!F"6!?+(*,!#,*=$!(##N($#>!*!
!"+!
b.24312!J8!-.24W!^5R8!Y.W7;2!CS8!R1G@!OJ!,#++(=>!OV6@1TF7!G;74612G!.34!G1_!2.@6;!""!
/;3T76/@!63!V;7W1<I2W;36/!V.2.G6@;64!H.GVG>!@<"*+($/!4$#.*$%+'#!!+(,%=$!&"*N"#!
&$$>!"$!
!"%!
b.24312!J8!R1G@!OJ!,#++%.=>!OV6@1!.<;3L!G6I763LG>!C(+"/("!#)&$!"%"$N"%"%>!"&!
!"'!
b.24312!J8!R1G@!OJ!,#++%I=>!OV6@1!.34!@B1!G/.71!;T!/;<V1@6@6;3>!-,.*/$%!,0!"(!
&6,%.#+,/$*3!5+,%,23$!"N*>!")!
!"*!
b.24312!J8!R1G@!OJ!,#++'=>!OV6@1>!>.**"/#!5+,%,23!!+,"(=$!]''#N]''%>!#+!
!#"!
b.24312!J8!R1G@!OJ!,^3!C21GG=>!b2113I1.24>!&6,%.#+,/>!##!
!#$!
b.24312!J8!R1G@!OJ8!EF/0763L!J!,#++%=>!E./@126;/63G8!GV6@1!.34!U62F713/1>!#%!
H*,(""9+/2'!,0!#)"!F,3$%!C,(+"#3!,0!D,/9,/!5!"*!$!"&#*N"&$&>!#&!
35!
!"!
b1U12G!S8!5;B.3!PK8!?.H213/1!\b8!OV2.@@!Eb8!5;13W1!Y8!P167!Q\!"#!$%!,#++&=>!]1N#!
1U.7F.@63L!V2;0.2W;@6/!GV1/61G>!4$#!F"6!?+(*,!#,*=$!($$N($*>!$!
!%!
b67!b8!O.I.@12NKFi;D!E8!C121DNE2;/.7!Z8!O67U.!P\8!?.@;221!J!,#++'=>!C7.G<64G!63!&!
@B1!.VB64!134;GW<I6;3@!EF/B312.!.VB646/;7.!H6@B!@B1!G<.771G@!L13;<1G>!J!'!
VFDD763L!1U;7F@6;3.2W!G@;2W>!E"/"!#*)$!"(N#&>!(!
!)!
b;7413T174!:8!R;1G1!5!,#++(=>!E6;7;LWeG!31_@!21U;7F@6;3>!4$#.*"!%%&,("#'=$!*!
$'*N$'*>!"+!
!""!
b26TT63!JO8!R1G@!OJ8!EF/0763L!J!,#++%=>!5;;V12.@6;3!.34!/;<V1@6@6;3!63!"#!
V.@B;L136/!I./@126.>!4$#.*"!%#),(++$=$!"+#%N"+#(>!"$!
!"%!
-./012!\8!5.23617!Q!,#++"=>!Q/;7;L6/.7!T6@31GG8!L13;<6/!6G7.34G!.34!I./@126.7!"&!
V.@B;L136/6@W>!J!S.2H636.3!U61H!;T!@B1!1U;7F@6;3!;T!<6/2;I1G>!"'!
>!&?5P!F"A,*#'!",&=$!$('N$)">!"(!
!")!
-.T@!]\P8!K6@@712!\Q8!Y2._712!E!,#++*=>!5;<V1@6@6;3!T.U;F2G!214F/14!/;G@!;T!"*!
V7.G<64G!@;!B;G@!I./@126.>!#)"!:C?&!-,.*/$%!#$!('"N('*>!#+!
!#"!
-.<67@;3!RS!,"*'%=>!YB1!L131@6/.7!1U;7F@6;3!;T!G;/6.7!I1B.U6;F2>!^!j!^^>!-,.*/$%!##!
,0!8)",*"#+($%!5+,%,23!*$!"N&#>!#$!
!#%!
36!
-.<67@;3!RS!,"*(+=>!O17T6GB!.34!GV6@1TF7!I1B.U6;F2!63!.3!1U;7F@6;3.2W!<;417>!"!
4$#.*"!""'$!"#")N"##+>!#!
!$!
-.2463!b!,"*')=>[email protected]!;T!@B1!/;<<;3G>!C(+"/("!!+"$!"#%$N"#%)>!%!
!&!
-.226G;3!P8!E2;H363L!?Q8!Z;G!K8!EF/0763L!J!,#++'=>!5;;V12.@6;3!.34!U62F713/1!63!'!
./F@1!CG1F4;<;3.G!.12FL63;G.!63T1/@6;3G>!5?>!5+,%,23!%,"=$!#">!(!
!)!
\;B3G;3!?\!,#++(=>!YB1!b13;<1!O@2601G!E./0[!YB1!QU;7F@6;3.2W!^<V;[email protected]/1!;T!*!
S1T13/1!JL.63G@!K;I671!Q71<13@G>!&6,%.#+,/$*3!5+,%,23!#%,$N%=$!"#"N"#*>!"+!
!""!
a17712!?8!];GG!ab!,"**)=>!O17T6GB!L131G[!.!L2113!I1.24!63!@B1!214!T621!.3@>!4$#.*"!"#!
#(%,''*$=$!&($N&(&>!"$!
!"%!
a3;@B1!-8!OB.B!C8!a2/<[email protected][email protected]!O!,"*)$=>!Y2.3GT12.I71!"&!
[email protected]/1!@;!/1T;@._6<18!/1T;_6@638!/1T.<.34;71!.34!/1TF2;_6<1!63!/7636/.7!"'!
6G;7.@1G!;T!a71IG6177.!V31F<;36.1!.34!O122.@6.!<.2/1G/13G>!:/0"(#+,/!!!,'=$!$"&N"(!
$"(>!")!
!"*!
a;I.W.GB6!^!,#++"=>!E1B.U6;2!;T!21G@26/@6;3N<;46T6/.@6;3!GWG@1<G!.G!G17T6GB!#+!
<;I671!171<13@G!.34!@B162!6<V./@!;3!L13;<1!1U;7F@6;3>!4.(%"+(!@(+9'!F"'"$*()!#"!
"(,")=$!$(%#N$(&'>!##!
!#$!
a;2;3.!]8!?1U63!E]!,"**$=>!CB.L1N<146.@14!G171/@6;3!T;2!21G@26/@6;3N#%!
<;46T6/.@6;3>!&6,%.#+,/!%*$!&'&N&(&>!#&!
37!
!"!
a;D7;H6/D!Ea8!OB6!a8!bF!gf8!AB7134;2T!S-8!Q.2B.2@!5J8!SF33W!bK!,#++'=>!#!
K;71/F7.2!I.G6G!T;2!/;3@2;7!;T!/;3MFL.@6;3!IW!I./@126.7!VB12;<;31!.34!63B6I6@;2!$!
V1V@641G>!?,%"(.%$*!<+(*,1+,%,23!+",%=$!*&)N*'*>!%!
!&!
aF27.34!5b8!5.3I./0!E8!E12L!Ab!,#++$=>!-;26D;[email protected][email protected][!.!/26@6/.7!U61H>!'!
H4@C!!)),"(=$!*'&)N*''#>!(!
!)!
?1412I12L!\8!Y.@F<!Q?!,"*%'=>!b131!21/;<I63.@6;3!63!Q>!/;76>!4$#.*"!!&'$!&&)>!*!
!"+!
?11!a8!?11!K8!OB63!\-8!?11!K-8!a.3L!O-8!C.20!J\!"#!$%!,#++'=>!C21U.713/1!;T!""!
C7.G<64N<146.@14!J<V5!kN?./@.<.G1G!63!&'()"*+()+$!(,%+!.34!Q%"1'+"%%$!"#!
A/".<,/+$"!63!a;21.>!?+(*,1+$%!7*.2!F"'+'#$/("!!","=$!%%N%*>!"$!
!"%!
?1B<.33!?8!P174<.3!KR8!P;G@12!a]!,#++)=>[email protected][email protected]<6GG6;3!/.3!63B6I6@!"&!
@B1!1U;7F@6;3!;T!.7@2F6G@6/!B17V63L>!@<"*+($/!4$#.*$%+'#!!*","=$!"#N#%>!"'!
!"(!
?1B<.33!?8!a17712!?!,#++'=>!YB1!1U;7F@6;3!;T!/;;V12.@6;3!.34!.7@2F6G<!N!.!")!
L1312.7!T2.<1H;20!.34!.!/7.GG6T6/.@6;3!;T!<;417G>!-,.*/$%!,0!&6,%.#+,/$*3!5+,%,23!"*!
!(,&=$!"$'&N"$('>!#+!
!#"!
?12.@!Q8!A/B<.3!-!,#++%=>!CG6NCB6[!1_V7;263L!@B1!;F@12!76<6@G!;T!I./@126.7!##!
VG1F4;L131G>!E"/,<"!*"'"$*()!!%,""=$!##($N##()>!#$!
!#%!
38!
?1U63!E][email protected]@!PK!,"*((=>!C2;I.I676@W!;[email protected]!/B6<126/!V7.G<64G!63!"!
[email protected]!V;VF7.@6;3G!;T!I./@126.>!C(+"/("!!(+$!#")N##+>!#!
!$!
?676!?:8!E26@@;3!:P8!P167!Q\!,#++(=>!YB1!V12G6G@13/1!;T!V.2.G6@6/!V7.G<64G>!E"/"#+('!%!
!**,"=$!$**N%+&>!&!
!'!
?6U12<;21!SK!,"**&=>[email protected]?./@.<.G1G!63!7.I;2.@;2W!.34!/7636/[email protected]/1>!(!
>%+/+($%!?+(*,1+,%,%,2,($%!F"6+"G'!',%=$!&&(N&)%>!)!
!*!
K.L3FG;3!]S!,#++(=>!-WV;@B1@6/.7!TF3/@6;3G!;T!@;_63N.3@6@;_63!GWG@1<G>!-,.*/$%!"+!
,0!5$("*+,%,23!!'(,"(=$!'+)*N'+*#>!""!
!"#!
K.2@631D!\?!,#++*=>!YB1!2;71!;[email protected]!13U62;3<13@G!63!@B1!1U;7F@6;3!;T!"$!
[email protected]/[email protected]@G!63!V.@B;L136/!I./@126.>!H*,(""9+/2'!,0!#)"!F,3$%!C,(+"#3!,0!"%!
D,/9,/!5!"*+$!#&#"N#&$+>!"&!
!"'!
K.2@l31D!\?!,#++)=>!J3@6I6;@6/G!.34!.3@6I6;@6/[email protected]/[email protected]!"(!
13U62;3<13@G>!C(+"/("!#"!$!$'&N$'(>!")!
!"*!
K.W!]K8!J3412G;3!]K!,"*(*=>!C;VF7.@6;3!I6;7;LW!;T!63T1/@6;FG!46G1.G1G[!C.2@!^^>!#+!
4$#.*"!"'),&(##=$!%&&N%'">!#"!
!##!
K;/B6DF06!J8!f.B.2.!a8!a;I.W.GB6!^8!^H.G.!f!,#++'=>!b131@6/!.446/@6;3[!G17T6GB!#$!
L131eG!G@2.@1LW!T;2!GW<I6;G6G!63!@B1!L13;<1>!E"/"#+('!!*"$!"$+*N"$#$>!#%!
!#&!
39!
:;LF162.!Y8!].3063!S\8!Y;F/B;3!K8!Y.4416!P8!E2;H3!OC8!];/B.!QC!,#++*=>!b131!"!
<;I676@W!426U1G!@B1!1U;7F@6;3!;T!I./@126.7!/;;V12.@6;3!.34!U62F713/1>!>.**"/#!#!
5+,%,23!!($!"')$N"'*">!$!
!%!
:;2<.3!J8!-.3G13!?-8!O;213G13!O\!,#++*=>!5;3MFL.@6U1!V7.G<64G[!U1GG17G!;T!@B1!&!
/;<<F3.7!L131!V;;7>!H)+%,',A)+($%!#*$/'$(#+,/'!,0!#)"!F,3$%!C,(+"#3!,0!D,/9,/!5!'!
#+%$!##(&N##)*>!(!
!)!
:F@6!KC8!?1V646!JJ8!C2.0.GB!]a8!O/B67V12;;2@!]J8!5.33;3!P5!,"*(*=>!QU6413/1!T;2!*!
36@2;L13!T6_.@6;3!,36T=!L131G!;3!6346L13;FG!]B6D;I6F<!V7.G<64G>!4$#.*"!"+!
"'",&($)=$!&$$N&$&>!""!
!"#!
AmE2613!JS8!:1H7.34!\R8!K67712!OP8!-;7<1G!]a8!O<6@B!-R8!P;2<.7!OE!,"*)%=>!"$!
OB6L.N7601!@;_63N/;3U12@63L!VB.L1G!T2;<!QG/B126/B6.!/;[email protected]!@B.@!/.FG1!"%!
B1<;22B.L6/!/;76@6G!;2!63T.3@671!46.22B1.>!C(+"/("!""+$!'*%N'*'>!"&!
!"'!
A/B<.3!-8!?.H213/1!\b8!b2;6G<.3!QJ!,#+++=>[email protected][email protected]!.34!@B1!"(!
3.@F21!;T!I./@126.7!633;U.@6;3>!4$#.*"!%)&$!#**N$+%>!")!
!"*!
A76U12!aK8!S1L3.3!C-8!-F3@12!KO8!K;2.3!:J!,#++*=>!E./@126;VB.L1G!Q3/;41!#+!
P./@;2G!]1dF6214!T;2!C2;@1/@6;3!63!.!OW<I6;@6/[email protected]<>!C(+"/("!#"&,&*%$=$!#"!
**#N**%>!##!
!#$!
A2L17!?Q8!526/0!P-5!,"*)+=>!O17T6GB!S:J[!YB1!F7@6<@1!V.2.G6@1>!4$#.*"!"'%$!'+%N#%!
'+(>!#&!
40!
!"!
C.F7GG;3!\!,#++#=>!KF7@671U1714!G171/@6;3!;3!V7.G<64!21V76/.@6;3>!E"/"#+('!#!
!+!,%=$!"$($N"$)%>!$!
!%!
CB67VV;3!J8!J271@!b8!\./;IW!bJ!,#++#=>!C7.G<64NS1@12<6314!J<[email protected]&!
?./@.<.G1G>!@/#+<+(*,1+$%!@2"/#'!$/9!>)"<,#)"*$A3!%+,"N""=>!'!
!(!
C;7D716@312!Q8!g1/B312!Q?8!]13312!R8!P2.@@1!]8!\.F0!E8!b!-!"#!$%!,"**(=>!Y2.K!;T!)!
V7.G<64!]"!/;3@2;[email protected]!L131!1_V21GG6;3!.G!.3!63@1L2.@14!/;3@2;7!171<13@!63!*!
.!/;<V71_!21LF7.@;2W!31@H;20>!?,%"(.%$*!<+(*,1+,%,23!"&,$=$!%*&N&+(>!"+!
!""!
].631W!CE8!].631W!a!,#++$=>!QU;7F@6;3!;T!/;;V12.@6;3!.34!/;3T76/@!63!"#!
1_V126<[email protected]!I./@126.7!V;VF7.@6;3G>!4$#.*"!%"&$!(#N(%>!"$!
!"%!
].3063!S\8!E.2LF<!a8!a;00;!-!,#++(=>[email protected]!;T!@B1!/;<<;3G!63!"&!
1U;7F@6;3.2W!I6;7;LW>!8*"/9'!+/!&(,%,23!$/9!&6,%.#+,/!""$!'%$N'&">!"'!
!"(!
].G/B14!^8!AI1212!Q!,"*)'=>!PT!/;76VB.L1G[!G@2F/@F2.7!.34!TF3/@6;3.7!")!
217.@6;3GB6VG>!?+(*,1+,%,%,2+($%!F"6+"G'!&),%=$!%+"N%#(>!"*!
!#+!
]6I;76NO.G/;!?8!E2;H3!OC8!Y.4416!P!,#++)=>!RBW!@1./Bn!YB1!1U;7F@6;3.2W!;26L63G!#"!
.34!1/;7;L6/.7/;3G1dF13/1G!;T!/;G@7W!63T;2<.@6;[email protected]>!^3[!Q@@;221!C!.34!##!
-FLB1G!S!,14G=!>,<<./+($#+,/!$<,/2!C,(+$%!P*2$/+'<'>!A_T;24!`36U12G6@W!#$!
C21GG>!#%!
!#&!
41!
]671W!KJ8!R12@D!\Q!,#++#=>!E./@126;/63G[!1U;7F@6;38!1/;7;LW!.34!.VV76/.@6;3>!"!
@//.$%!F"6+"G!,0!?+(*,1+,%,23!&+$!""(N"$(>!#!
!$!
];FGG1@!P!,#++%=>!E"/"#+(!'#*.(#.*"!$/9!'"%"(#+,/!+/!'.19+6+9"9!A,A.$%#+,/'>!%!
C263/1@;3!`36U12G6@W!C21GG[!C263/1@;3>!&!
!'!
O./BG!\?8!KF17712!`b8!R67/;_!YC8!EF77!\\!,#++%=>!YB1!1U;7F@6;3!;T!/;;V12.@6;3>!8)"!(!
M.$*#"*%3!F"6+"G!,0!5+,%,23!*(,#=$!"$$N"'+>!)!
!*!
O./BG!\?8!O6<<G!Q?!,#++(=>[email protected]!@;!<[email protected]<!I21.04;H3>!8*"/9'!+/!"+!
&(,%,23!$/9!&6,%.#+,/!"!,"+=$!&)&N&*#>!""!
!"#!
O.3/B1D!E8!E21GG;77612!C8!`24./6!K5!,#++)=>!Q_V;2@14!V2;@163G!63!V2;I6;@6/!"$!
I./@126.[!J4B1G6;3!@;!63@[email protected]!GF2T./1G8!B;G@!6<<F3;<;N!4F7.@6;3!.34!"%!
<;71/F7.2!/2;[email protected]!H6@B!@B1!B;G@>!B&?C!:<<./,%,23!$/9!?"9+($%!"&!
?+(*,1+,%,23!&%$!"X"(>!"'!
!"(!
O7.@12!P]8!E.671W!K\8!Y1@@!J\8!YF2312!O?!,#++)=>!C2;L21GG!@;[email protected]!")!
@B1!T.@1!;T!V7.G<64G!63!I./@126.7!/;<<F36@61G>!B&?C!?+(*,1+,%!&(,%,23!++$!$N"$>!"*!
!#+!
O7.@12!O58!b;74<.3!EO8!b;;[email protected]!\58!P.22.34!Oa8!:1G@12!QR!"#!$%!#"!
,#++*=>!b13;<1!O1dF13/1G!;T!YB211!JL2;I./@126F<!E6;U.2G!-17V!Q7F/64.@1!@B1!##!
QU;7F@6;3!;T!KF7@6/B2;<;G;<1!b13;<1G!63!E./@126.>!-!5$(#"*+,%!!(!,)=$!#&+"N#$!
#&"">!#%!
!#&!
42!
O<6@B!\!,#++"=>!YB1!G;/6.7!1U;7F@6;3!;T!I./@126.7!V.@B;L131G6G>!H*,(""9+/2'!,0!#)"!"!
F,3$%!C,(+"#3!,0!D,/9,/!C"*+"'!5K5+,%,2+($%!C(+"/("'!"+'$!'"N'*>!#!
!$!
O;213G13!O\8!E.671W!K8!-.3G13!?-8!a2;12!:8!RF12@D!O!,#++&=>!O@F4W63L!V7.G<64!%!
B;26D;[email protected][email protected]!63!G6@F[!.!/26@6/.7!21U61H>!4$#!F"6!?+(*,!#,*=$!(++N("+>!&!
!'!
[email protected]@!PK8!?1U63!E]!,"*((=>!YB1!V;VF7.@6;3!I6;7;LW!;T!I./@126.7!V7.G<64G[!.!(!
V26;26!/;346@6;3G!T;2!@B1!1_6G@13/1!;T!/;3MFL.@6;[email protected]<6@@14!T./@;2G>!)!
E"/"#+('-'*$!#+*N##)>!*!
!"+!
YB;<.G!5K8!:617G13!aK!,#++&=>!K1/B.36G<G!;T8!.34!E.22612G!@;8!-;26D;[email protected]!b131!""!
Y2.3GT12!I1@H113!E./@126.>!4$#!F"6!?+(*,!#,*=$!(""N(#">!"#!
!"$!
Y;F/B;3!K8!-;141!58!Y13.677;3!A8!E.2I1!Z28!E.126GHW7!O8!E641@!C!"#!$%!,#++*=>!"%!
A2L.36G14!b13;<1!SW3.<6/G!63!@B1!&'()"*+()+$!(,%+!OV1/61G!]1GF7@G!63!-6LB7W!"&!
S6U12G1!J4.V@6U1!C.@BG>!HD,C!E"/"#!&,"=$!1"+++$%%>!"'!
!"(!
YF2312!CQ!,#++%=>!CB13;@WV6/!V7.G@6/6@W!63!I./@126.7!V7.G<64G>!E"/"#+('!!+*,"=$!")!
*N#+>!"*!
!#+!
YF2312!CQ8!5;;V12!ZO8!?13G06!]Q!,"**)=>!Y2.41;TT!I1@H113!B;26D;[email protected]!.34!#"!
U12@6/.7!<;41G!;[email protected]<6GG6;3!63!I./@126.7!V7.G<64G>!&6,%.#+,/!&",#=$!$"&N$#*>!##!
!#$!
43!
U.3!412!C7;1L!\]!,#++&=>!]1LF7.@6;3!;T!E./@126;/63!C2;4F/@6;3!63!C#*"A#,(,((.'!"!
<.#$/'!IW!@B1!dF;2F<NG13G63L!GWG@1<!21dF6214!T;2!41U17;V<13@!;T!L131@6/!#!
/;<V1@13/1>!-,.*/$%!,0!5$(#"*+,%,23!!'*,"#=$!$*)+N$*)*>!$!
!%!
Z.3!K1741213!?8!S1!E.G@!KO!,#++*=>!E./@126.7!@;_63N.3@6@;_63!GWG@13G[!<;[email protected]!&!
G17T6GB!13@6@61Gn!HD,C!E"/"#+('!&,$=$!1"+++%$(>!'!
!(!
R.L312!J!,#++*=>!Y2.3GV;G.I71!171<13@G!.G!L13;<6/!46G1.G1G>!?,%"(.%$*!)!
5+,'3'#"<'!&,"=$!$#N$&>!*!
!"+!
R.74;2!Ka8!K10.7.3;G!\!,"**'=>!?WG;L136/!/;3U12G6;3!IW!.!T67.<13@;FG!VB.L1!""!
13/;463L!/B;712.!@;_63>!C(+"/("!"*"$!"*"+N"*"%>!"#!
!"$!
R1G@!OJ8!S6LL71!OC8!EF/0763L!J8!b.24312!J8!b26TT63!JO!,#++(.=>!YB1!G;/6.7!76U1G!;T!"%!
<6/2;I1G>!@//.$%!F"6+"G!,0!&(,%,23R!&6,%.#+,/!$/9!C3'#"<$#+('!#'$!&$N((>!"&!
!"'!
R1G@!OJ8!b26TT63!JO8!b.24312!J!,#++(I=>!QU;7F@6;3.2W!1_V7.3.@6;3G!T;2!"(!
/;;V12.@6;3>!>.**"/#!5+,%,23!!*,"'=$!]''"N]'(#>!")!
!"*!
R1G@!OJ8!b26TT63!JO8!b.24312!J!,#++(/=>!O;/6.7!G1<.3@6/G[!.7@2F6G<8!/;;V12.@6;38!#+!
<[email protected]<8!G@2;3L!21/6V2;/6@W!.34!L2;FV!G171/@6;3>!-,.*/$%!,0!&6,%.#+,/$*3!#"!
5+,%,23!")$!%"&N%$#>!##!
!#$!
R1G@!OJ8!b26TT63!JO8!b.24312!J8!S6LL71!OC!,#++'=>!O;/6.7!1U;7F@6;3!@B1;2W!T;2!#%!
<6/2;;2L.36G<G>!4$#.*"!F"6+"G'!?+(*,1+,%,23!%$!&*(N'+(>!#&!
44!
!"!
RB6@1!5Q8!R63.3G!O5!,#++(=>!5177N/177!/;<<F36/.@6;3!63!@B1!V7.3@!V.@B;L13!#!
@2*,1$(#"*+.<!#.<"0$(+"/'>!H)+%,',A)+($%!8*$/'$(#+,/'!,0!#)"!F,3$%!C,(+"#3!,0!$!
D,/9,/!C"*+"'!5K5+,%,2+($%!C(+"/("'!#+"$!""$&N""%)>!%!
!&!
h.U612!\E8!P;G@12!a]!,#++(=>!5;;V12.@6;3!.34!/;3T76/@!63!<6/2;I6.7!I6;T67<G>!'!
H*,(""9+/2'!,0!#)"!4$#+,/$%!@($9"<3!,0!C(+"/("'!!)%$!)('N))">!(!
!)!
f.@1G!5K8!OB.H!S\8!];1!J\8!R;;7B;FG1!KQ\8!J<W1G!ObE!,#++'=>!Q3B.3/1<13@!;T!*!
I./@126.7!/;<V1@6@6U1!T6@31GG!IW!.V2.<W/[email protected]/1!V7.G<64G!T2;<!3;3N"+!
V.@B;L136/!&'()"*+()+$!(,%+>!5+,%,23!D"##"*'!",$=$!%'$N%'&>!""!
!"#!
g63412!:S8!?1412I12L!\!,"*&#=>!b131@6/!Q_/B.3L1!63!O.7<;3177.>!-,.*/$%!,0!"$!
5$(#"*+,%,23!+%$!'(*N'**>!"%!
!"&!
gFV.3!\8!KF@B!Y]8!S2.V12!A8!g.<I2WG06!C!,#+++=>[email protected]!;T!S:J!T2;<!"'!
JL2;I./@126F<!@F<1T./613G!63@;!V7.3@G[!.!T1.G@!;T!TF34.<[email protected]!63G6LB@G>!8)"!H%$/#!"(!
-,.*/$%!"#$!""N#)>!")!
"*!
45!
Figure Legends "!
Figure 1. An illustration of the nested nature of plasmids and their bacterial hosts. #!
Different sub-populations may harbour different densities of plasmids. If a gene $!
involved in cooperation is carried on a plasmid, this will affect the relatedness %!
between individuals in a patch (due to relatedness being measured at a focal locus, in &!
this case on a plasmid). Thus, sub-populations A and D have high plasmid-'!
relatedness, while sub-populations C and B have low plasmid relatedness. Relatedness (!
is influenced by local cell density, and migration between sub-populations and, in the )!
case of loci on plasmids, relatedness changed due to the degree of horizontal gene *!
transfer (see box 2 and figure 3 for details). "+!
""!
Figure 2. General classification of social traits carried by MGEs and their effect on "#!
the bearer or a social neighbour. x represents the impact of the trait on the fitness of "$!
the host bacteria, while y represents the impact of the trait on the fitness of the host "%!
bacteria’s neighbours. "&!
"'!
Figure 3. The effect of horizontal gene transfer rate ! on genetic relatedness between "(!
individuals within a patch, measured at loci with transfer rate !. The equation follows ")!
form that described in box 2, where m=0.1, N=50 and s=0. In the absence of any HGT "*!
(i.e. !=0), for the parameters chosen, r=0.079 (shown by the dotted line). #+!
#"!
Sub population or group Dispersal
and competition
bacteria
plasmid
A
B
C
D
chromosome
!"#$%&'(&)'$*+&,(-,.,-/*+&0!1&
!"#$%&'(&(#,234'/5&0"1&
6+%5/,78&
9:,%#& 9#+;73(#77&
</%/*+&=#(#;%&
>?&@*5*7,A$& </%/*+,7A$&!&
B#+:,(2&C&
D"&B*58,(2&
0.2 0.4 0.6 0.8 1.0Horizontal gene transfer Β
0.2
0.4
0.6
0.8
1.0
Relatedness R�
Rate of transfer (!)"
Rela
tedness
(R*)"
!"#$%&$'()'*+,'