Welcome to a Scintilla of Playful Musings

Welcome to my new blog, noos anakainisis, translated literally as mind renewal. The primary obsessions are neuroscience, computation, information, structure, form, art and history of science. Some environmental, political, and technological developments will also be included.

I hope your neurons are sufficiently stimulated...

Tuesday, September 7, 2010

The Evolution of Spite (and Altruism)

Behaviors that decrease the relative fitness of the actor--and also either benefit (altruism) or harm (spite) other individuals--are difficult to reconcile with natural selection and maximization of individual fitness. The paragon of altruism is the sterile worker caste within eusocial insect colonies, which help rear the offspring of their queen, or the slime mold cells that altruistically give up their own survival to become the nonviable stalk of a fruiting body, helping other cells to disperse in the form of spores. These behaviors reduce the reproductive success of the altruist--so why doesn't natural selection weed out the genes responsible for such behaviors?

Hamilton showed that genes can spread not only through their direct impact on their own transmission, but also through their indirect impact on the transmission of copies present in other individuals. He introduced the theoretical concept of inclusive fitness--Hamilton's Rule--which states that a trait will be favored by selection when rb-c>0, where c is the fitness cost to the actor, b is the fitness benefit to the recipient and r is their genetic relatedness. Consequently, altruistic behaviors are favored if the benefits are directed toward other individuals who share genes for altruism.

Eusociality, depending on how it is defined, has evolved 3-11 times in Hymenoptera (ants, bees, wasps), termites, thrips, aphids, spiders, beetles, shrimps and mole rats. A crucial parameter necessary for the evolution of eusociality is strict, one-time monogamy (which has been shown as the ancestral state in all independent origins of eusociality studied), in which females only mate one male in their entire life. This monogamy leads to a potential worker being equally related (r=0.5) to her own offspring and to the offspring of her mother (siblings). In this case, any small efficiency benefit for rearing siblings over their own offspring (b/c>1) will favor eusociality (such benefits include life insurance=helpers completing parental care after the death of the mother, as well as fortress defense=help use or defend a food source when opportunities for successful migration are low). Later in evolutionary trajectories of eusocial animals, once the workers have lost the ability to mate and realize full reproductive potential themselves and generally have specialized to a division of labor that gives a substantial b/c (a large efficiency benefit for sibling-rearing since siblings are less related to the individual than their own offspring would be), some queens develop the ability/behavior to mate with multiple males.

Spiteful behaviors would be favored--i.e. rb-c>0 is satisfied--if c is positive (which is costly to the actor) and b is negative (which is costly to the primary recipient of the spiteful behavior), only if relatedness between the actor and recipient, r, is negative (negative relatedness is when the recipient is less related to the actor than expected by chance). The indirect fitness of spite is that secondary recipients, more closely related to the actor than the primary recipient, experience reduced competition from the primary recipient harmed by the spiteful behavior. Spite is therefore altruism to the secondary recipients: harming an individual is favored if it provides a benefit closer relatives.

Some confusion about spite arose due to certain behaviors only being evaluated with respect to direct fitness over the short term rather than over the lifetime of the actor; these include bird siblicide at neighboring nests and fish egg cannibalism (decreased competition for resources for the actor and/or actor's offspring), mammalian infacticide, especially of juvenile males (decreases competition for offspring or mates) and human punishment/rejection of low offers in economic games (increased cooperation over the long term). All of these examples are selfish behaviors that are costly to the recipient but provide a benefit to the actor (c<0). The specific conditions required to favor evolutionary spite, population structures in which harming non-relatives is an efficient way of helping relatives, may be rare in general and unlikely in humans and other primates.

An example of spite can be found in the polyembryonic parasitoid wasps.  A female wasp lays eggs on moth caterpillars, after which the wasp eggs divide asexually into many larvae and consume the growing caterpillar from the inside. Most larvae develop normally, but a fraction become soldier morphs. Developing as a soldier is costly to the actor (they are sterile) and costly to the primary recipient (soldiers seek out and kill larvae that developed from the other eggs within the host), however it is beneficial to the soldier's clone-mates that developed from the same egg, freeing up resources (the caterpillar body) for their consumption.

From a theoretical perspective, spite is plausible if there is large variance in relatedness between competitors, kin discrimination (with harming behaviors aimed at individuals to whom the actor is relatively unrelated), and strong local competition so that harming the primary recipient provides appreciable benefits to the secondary recipients.  Local competition for resources typically selects for spite and against altruism; altruistic traits show a positive, monotonic relationship to relatedness, whereas spiteful traits show a domed relationship; kin discrimination is key for spite, whereas altruism can often evolve without kin discrimination when limited dispersal keeps relatives together.

As Hamilton has pointed out, the indirect fitness benefits derived from altruism and spite require genetic relatedness per se, not kinship (ie. genetic relatedness at the altruism locus, not geneaological relationship over the whole genome). This can be accomplished in two ways: a gene or set of tightly-linked genes that both cause the cooperative behavior and cause cooperators to associate (coined "greenbeards" by Dawkins) or by geneaological kinship. In the slime mold, Dictyostelium discoideum, individuals with the csa gene adhere to each other in aggregation streams and cooperatively form fruiting bodies while excluding noncarriers of the gene. A spiteful greenbeard in fire ants, Solenopsis invicta, is the b allele of the Gp-9 gene, which enables workers to use oder to determine whether prospective queens also carry this allele, dismembering them if they do not.

There are four categories of greenbeards: altruistic and always expressed (obligate), altruistic and only expressed in response to presence of greenbeard in others (facultative), spiteful and obligate, spiteful and facultative.  For all cases except altruistic facultative, the greenbeard is selected against at low frequencies and only favored when it has established itself to a certain frequency. Population structure can solve this problem by keeping individuals with greenbeards together. Some models for altruism in humans implicitly invoke greenbeard mechanisms (suggesting altruistic individuals differ from non-altruistic individuals in some observable characteristic like smiling or tendency for punishment), which is only true if the greenbeard mechanism is encoded by the same gene or closely linked genes as those that lead to the altruism, otherwise falsebeards could too easily arise and the altruism (and its detection) would not be evolutionarily stable.

Microbes are ideal model organisms to look for new greenbeards because their asexual growth leads to extreme population structuring, the genotype is relatively simply linked to the phenotype and this simplicity may prevent decoupling between the greenbeards and falsebeards (cheats that displayed the signal without also performing the behavior), and genetic knockouts can be designed to aid in the detection of greenbeards.  

Another example of spite is the costly production and release of antimicrobial bacteriocins, toxins that can kill unrelated strains of the same species that lack the specific immunity gene. In some cases, cell death is required to release the bacteriocins into the environment, so it is clearly costly to the actor. The bacteriocin production genes are genetically linked to the immunity genes so that close relatives both produce it and are immune to it. When one bacteria does release its bacteriocin, it will thus only kill non-relatives and free up resources for clone-mates.

In the Hawlena study, two natural populations of Xenorhabdus bacteria are carried by entomopathogenic nematodes, dispersing over a range of a few metres within these symbiotic hosts, and use bacteriocins as weapons. The authors found that genetic relatedness decreased and the probability of bacteriocin-mediated (i.e. spiteful) interactions increased with spatial distance between isolates. Measurements were taken at a scale ranging from 1 to 120 metres. Whilst this work has only been done on a relatively small scale and in one system, it is clearly important to test theoretical results with real systems and, fortunately, in this case, the experimental results support the theory.

Hamilton WD. (1963) The Evolution of Altruistic Behavior.  American Naturalist. 97:354-6.

Wloch-Salamon DM, Geria D, Hoekstra RF, deVisser JAGM. (2008) Effect of dispersal and nutrient availability on the competitive ability of toxin-producing yeast. Proc R Soc Lond B. 275:535-41

Hawlena H, Bashey F, Lively CM. (2010) The Evolution of Spite: Population Structure and Bacteriocin-Mediated Antagonism in Two Natural Populations of Xenorhabdus Bacteria. Evolution.

West SA, Gardner A. (2010) Altruism, Spite, and Greenbeards. Science. 327:1341-1344.

Related Posts by Categories


Post a Comment