National Geographic

‘Friendly’ genes are more likely to be passed around


Invisible to the naked eye, a frenetic marketplace buzzes all around you. The customers are bacteria and they are trading in genes, swapping them between individual cells as easily as humans swap presents or phone numbers. Some of the trades allow bacteria to cope with new sources of food. Others are more like arms dealing, with cells exchanging genes that allow them to beat antibiotics, or weapons that bestow the ability to cause disease.

These swaps are pervasive. At least an eighth of the genome of E.coli, a commonly studied species, has been borrowed from other bacteria. But this leads to an interesting puzzle. Genes don’t work in a vacuum; they interact with one another in a tangled web of partnerships. Some depend on other genes to switch them on or hold them back. Some encode proteins that only work in tandem. So how does a gene that finds itself in a foreign land manage to do anything, let alone play an active role in the evolution of its new host?

The commonly cited answer is that genes are more likely to be transferred if they’re loners. The ones that keep to themselves, and depend less on their relations with other genes, are most likely to flourish in a new setting. Think about a car. If you take the seats or axles from a Mini and stick them in a Ferrari, you’re in for a poor ride – they’re too closely tied to the vehicle’s other parts. However, you could swap their rear-view mirrors without much difficulty.

This idea – the so-called “complexity hypothesis” – makes sense, but Israeli scientists Uri Gophna and Yanay Ofran think that it’s wrong. According to them, it’s actually the “friendly” genes – the ones that interact most with others – that are more likely to be transferred successfully. Forget the car. An office worker makes for a better analogy. A loner moving into a new workplace would keep to themselves without integrating into their new environment. But a more sociable employee would quickly start to fit in, making new contacts and forging new relationships. The same is apparently true for bacterial genes.

Gophna and Ofran considered the entire genome of E.coli. For each gene, they worked out how “friendly” their proteins are with other proteins, and how likely they are to be transferred to different species.

All proteins are made up of chains of amino acids and some are essential for interacting with other partners. For each E.coli protein, Gophna and Ofran identified and counted these ‘sticky’ amino acids. They called this number, expressed as a percentage of the total amino acids, the RIP or “relative interaction potential”. It’s a bit like counting the number of USB sockets in a computer – it gives you an idea of how many other devices can be plugged into it.

Next, Gophna and Ofran compared E.coli’s proteins to their counterparts in other bacteria. If bacteria only passed down their genes from mother to daughter, the evolutionary histories of all E.coli’s proteins would be much the same. If you created a family tree for one protein, linking its different versions from different bacterial species, it would look pretty much like the family tree for any other protein. But if genes are exchanged between species (and they are), this tidy comparison breaks down. So comparing the discrepancies between different trees gives you an idea about which proteins have been swapped most often.

At first, Gophna and Ofran’s results seemed to support the complexity hypothesis. Proteins that interact more with other proteins were less prone to inter-species swaps. But there’s a big flaw in this conclusion – the data only looked at the proteins after their genes had jumped into E.coli. They say nothing about what the genes were like before they jumped.

This is important because after a jump, genes change. Imagine our office worker again. If they’re reclusive, it might be that they were always antisocial. Alternatively, they might have spent so long in the same place that their social streak has gone. The same is true for jumping genes. As their proteins find new niches and roles for themselves, the amino acids that were essential for interactions in their past lives may warp and degenerate. Their current “friendliness” doesn’t necessarily reflect on what they were doing in their original species. And this historical perspective is essential for working out which genes are most likely to jump in the first place.

To cut through this confusion, Gophna and Ofran focused on the proteins that E.coli gained from other species, and divided them into groups according to when they were imported. The earliest ones would have had ample time to integrate into their new genome. The most recent ones are probably closest in form to their original versions in other species.

These recent acquisitions were clearly friendlier than the more ancient ones, and indeed than the rest of E.coli’s proteins. So much for the complexity hypothesis – Gophna and Ofran’s calculations suggest that when jumping genes land in a new host, they’re more likely to gain a foothold if their proteins interact with lots of others.

Of course, this isn’t a universal rule. Many genes are transferred as a package rather than as solo agents, and others are quite capable of going about their duties without any partners. But these are in the minority. At least in bacteria, most jumping genes are social ones that end up in new hosts without their usual partners. Those that have the greatest potential for interacting with others are more likely to fit in.

This isn’t just a case of the incoming proteins finding counterparts of their original partners in their new hosts. Gophna and Ofran calculated that even if the new partners were more than 90% identical to the old ones, the protein still probably wouldn’t pair up well with it. Instead, the success of “friendly” jumping genes lies in being able to find new partners to interact with.

Previous studies support this idea. Far from being isolated, the genes that E.coli gets from other species tend to interact with other well-connected genes. They find themselves at the centre of busy hubs, rather than stranded on the edges of the networks. Not only that, but they have a habit of working together with the core genes – the essential ones that the bacterium cannot do without.

Reference: PNAS

Image by Alan Cann

More on jumping genes:

There is 1 Comment. Add Yours.

  1. Dan Gaston
    December 14, 2010

    Although I haven’t had a chance to read the paper yet, since it is not available until later in the week, I ahd a few brief thoughts on this issue after discussing with a colleague. One is the claim that transfered proteins with high connectivity are unlikely to “fit” well with their xenologous counterparts after a transfer event even if they are at the 90% identity level. When talking about protein-protein interactions, global sequence divergence is not going to be an appropriate measure, that 10% sequence divergence may be located entirely within surface loops that are uninvolved, even allosterically, in the protein-protein interaction. It will be primarily the sequence (and structural) conservation in the binding sites that will be important. I have worked with proteins involved in large protein complexes that are hugely divergent that still function well in complementation studies for instance, because the proper interaction sites where mostly conserved.

    Which is my second point/question. These more recent HGT events, what sort of evolutionary distance were they transfered over? Within strain or within Phylum transfers of proteins with high connectivity seem much more likely to me than HGT events of highly connected proteins over large evolutionary distances in the tree.

    For those interested Cohen, Gophna, and Pupko have an article out concerning gene family acquisition by HGT out in MBE Advance Access here

Add Your Comments

All fields required.

Related Posts