Harmit Malik studies the causes and consequences of genetic conflicts that take place between different genomes (e.g., host-virus interactions, mitochondrial conflicts with nuclear genomes) or between components of the same genome (e.g., chromosomal competition at centromeric regions). He is interested in understanding these "molecular arms races" and how they drive recurrent genetic innovation, from the perspective of both evolutionary biology and human disease.
Eukaryotic cells harbor a conglomerate of different genetic entities, each locked in conflict with other genetic entities for evolutionary dominance, often within the same genome. Still other groups of genes have been shaped by evolutionarily ancient conflicts with infectious pathogens, often emerging as victors from such conflicts with compromised housekeeping functions in the cell. My lab is interested in understanding how eukaryotic genetics and epigenetics have been shaped by such events. We study causes, mechanisms, and consequences of two forms of genetic conflict: extrinsic (between genomes) conflicts that shape genes involved in host-pathogen interactions, and intrinsic (within genome) conflicts that shape eukaryotic genome architecture.
Evolution as a Tool to Describe and Discover Antiviral Strategies
In the past decade, virologists have revealed a novel arm of intracellular, cell-autonomous immunity that mammalian cells mount against a variety of viral infections. For instance, the TRIM5alpha gene was discovered in a functional screen because it confers resistance to HIV infection in rhesus macaques but not in humans. In collaboration with Michael Emerman and Adam Geballe (Fred Hutchinson Cancer Research Center), we have used an evolutionary approach that identifies potential antiviral genes based on evolutionary signatures that suggest that they have been locked in antagonistic conflict with viruses throughout primate evolution. These conflicts result in higher than expected rates of amino acid changes that are fixed by selection (positive selection). Despite their discovery, arising from their activity against young, extant viruses such as HIV-1, our analyses of the evolutionary histories of intracellular immunity genes have revealed that they have acted as antiviral genes throughout mammalian evolution. Several immunity genes, such as TRIM5alpha and APOBEC3G, represent some of the fastest-evolving genes in the human genome.
Figure 1: Evolution at host-virus interfaces. When host antiviral proteins specifically recognize a viral factor to mediate their antiviral effect, this kind of antagonism forces viruses to adapt at their interaction surface, which in turn drives host adaptation. Individual changes in hotspot residues at the molecular interface (shown in red) can determine whether the host or virus is winning. Signatures of this rapid evolution (positive selection) can thus identify the tempos and mode of host-virus interactions. Adapted from Daugherty, M.D., and Malik, H.S. 2012. Annual Review of Genetics 46:677–700. Figure 2: The centromere-drive model. Expansion of a satellite that binds Cid provides more microtubule attachment sites. This stronger centromere drives female meiosis, but also leads to increased nondisjunction. A mutation in Cid that alters sequence specificity leads to more extensive binding of the weaker centromere, providing more microtubule attachment sites. This restores meiotic balance and alleviates nondisjunction. From Bayes, J.J., and Malik, H.S. 2008 eLS doi:10.1002/9780470015902.a0020827. © 2008 John Wiley & Sons, Ltd. All rights reserved.
Malik Research Abstract Slideshow
Figure 1: Evolution at host-virus interfaces. When host antiviral proteins specifically recognize a viral factor to mediate their antiviral effect, this kind of antagonism forces viruses to adapt at their interaction surface, which in turn drives host adaptation. Individual changes in hotspot residues at the molecular interface (shown in red) can determine whether the host or virus is winning. Signatures of this rapid evolution (positive selection) can thus identify the tempos and mode of host-virus interactions.
Adapted from Daugherty, M.D., and Malik, H.S. 2012. Annual Review of Genetics 46:677–700.
Figure 2: The centromere-drive model. Expansion of a satellite that binds Cid provides more microtubule attachment sites. This stronger centromere drives female meiosis, but also leads to increased nondisjunction. A mutation in Cid that alters sequence specificity leads to more extensive binding of the weaker centromere, providing more microtubule attachment sites. This restores meiotic balance and alleviates nondisjunction.
From Bayes, J.J., and Malik, H.S. 2008 eLS doi:10.1002/9780470015902.a0020827. © 2008 John Wiley & Sons, Ltd. All rights reserved.
We have used the signature of positive selection to identify the amino acid residues in antiviral proteins that are responsible for specific recognition of viral components. This was one of the first instances where positive selection guided the functional analysis of host-pathogen interfaces. For example, the signature of positive selection in TRIM5alpha identified a patch of only a few amino acids that turned out to be responsible for determining the recognition of specific retroviral capsids, including HIV. More recently, we showed that such an evolution-guided approach could also reveal the interaction interface of the broadly acting MxA antiviral protein, which had eluded investigations that used more traditional virology and biochemical approaches.
Our analyses have revealed that even single–amino acid changes could determine the outcome of such host-virus conflicts. We postulated that viruses much older than those in the present day have driven selection for our current antiviral specificities. This has led us to propose an alternate approach of "indirect paleovirology," i.e., inferring the presence and action of ancient viruses by virtue of the evolutionary episodes of selection they drive in host antiviral genes. Together with the identification of fossilized imprints of ancient viruses in animal genomes, these reveal an ancient tapestry of viral infections throughout animal evolution.
Our survey of polymorphisms in antiviral genes in human populations also showed that retention of antiviral defense could be quite labile. We found alleles associated with impaired antiviral activity in both TRIM5alpha and APOBEC3H in certain human populations, suggesting that demography and relaxed selective pressures may have led to a significant reduction in the antiviral repertoire in many human populations. Thus, evolution can provide a means for identifying potential antiviral genes, revealing the functional sites of host-virus antagonism and ancient viruses themselves, and understanding differences in human susceptibility to infectious diseases. We are interested in expanding this host-virus paradigm to that between host genomes and their resident retroelements.
Viral Mimicry and "Accordions"
One highly successful strategy employed by both viral and bacterial pathogens to subvert host cell machinery for their own purposes is mimicry. For instance, the K3L gene from a lineage of poxviruses (that includes smallpox) resulted from the acquisition of a host translation gene, eIF2alpha. K3L aids poxviruses in defeating the antiviral response mounted by the PKR (protein kinase R) gene. PKR phosphorylates eIF2alpha on detecting viral infection to block protein production. We discovered that positive selection of PKR can allow it to overcome the challenge imposed by viral mimicry. This positive selection is most pronounced at the interaction surface between PKR and its substrate, eIF2alpha. This is intriguing because eIF2alpha is highly conserved in eukaryotes, whereas PKR is among the fastest-evolving genes in primate genomes. Our analyses revealed that host genes can employ evolutionary strategies to overcome mimicry by pathogens.
We also recently showed that when confronted with a PKR version that it cannot inhibit, poxviruses often undergo dramatic accordion-like expansions of their genome, especially on the K3L locus, providing temporary relief against PKR via mass action. This expanded pool of K3L genes also increases the probability of acquiring amino acid adaptations such that K3L can single-handedly defeat PKR, following which the "accordion" can collapse. We are interested in exploring whether such gene accordions drive both viral and host adaptations in genetic conflicts.
Genetic Conflict as a Basis for Centromere Complexity (and Speciation?)
Centromeres are the DNA sites of attachment of microtubules that orchestrate the orderly movement of chromosomes during cell division. However, there is a range of complexity and rapid evolution of centromeric DNA in various eukaryotes that is hard to explain for a process that seems so central to eukaryotic replication. Steven Henikoff (HHMI, Fred Hutchinson Cancer Research Center) and I proposed the "centromere-drive" hypothesis to resolve this apparent paradox. We proposed that centromeres actively compete with each other for evolutionary survival. In the process of female meiosis, only one of four meiotic chromosomes "wins," i.e., is chosen in the oocyte in both plants and animals. The other three chromosomes are eliminated. We postulated that centromeres, by virtue of their microtubule attachment, are under direct selection for inclusion in the egg, which provides a robust impetus for evolutionary changes and expansions at the centromere.
Although success in female meiosis would be a significant advantage for the winning centromere, the ensuing imbalances in centromere strength could result in either sex-ratio distortion or male sterility. Thus, other genes of the genome would be predicted to evolve to suppress these evolutionary costs of the driving centromere. Consistent with this prediction, we found that ancient centromeric proteins evolve under positive selection, but also that young proteins are often recurrently recruited for centromere function, rapidly becoming essential by virtue of their adopted roles in chromosome segregation. Thus, although chromosome segregation is essential and despite the fact that centromeric proteins and DNA are both required for this process, they are nonetheless locked in an evolutionary conflict playing out with dynamics similar to those of host-virus interactions. Current work is aimed at determining whether incompatibilities between centromeric components can provide a reproductive barrier between closely related species.
Grants from the National Institutes of Health and the Mathers Foundation provided partial support for these projects.
As of February 22, 2016