The route to rewiring – technical background

It’s quite common for scientific discoveries to be made by chance. Our chance discovery was that we accidentally left experimental samples out for longer than needed and in doing so something remarkable happened. Here we discuss our findings that have recently been published in Science []. Our aim here is to give the reader extra background that we couldn’t fit into the paper, to provide some insight to the historical background of why we were studying the evolution of gene regulation and to also show the step-wise key experimental steps that took place. We have developed two versions of the discussion – a lay version and a technical version.

Our story starts 11 years ago, when Rob Jackson and colleagues were examining the basis of bacterial gene regulation. We were using a common soil bacterium called Pseudomonas fluorescens. When looking for P. fluorescens mutants that could turn on their genes for making the extracellular polysaccharide, cellulose, we found several mutants with an inactivated gene called fleQ. FleQ is a master regulator of the bacterial flagellum, a propeller-like appendage that helps it to move through liquids and gels. The job of FleQ is to bind to DNA at specific enhancer sites in the vicinity of flagellum genes and bind RpoN to activate transcription. When we mutated fleQ the flagellum genes could not be expressed and the cell was thus unable to synthesise a flagellum. As a consequence, we observed [] that these mutant bacteria were no longer able to swim, but could still move slowly over agar surfaces, exhibiting an unusual spreading pattern called spidery-spreading. We found that the basis of this spreading was due to a surfactant called viscosin – mutation of the genes for making both viscosin and a flagellum (double mutant strain) resulted in the bacteria being immotile [;jsessionid=FACAEA8FFAAA1CBF36F50E1A5291AB85.f01t03].

Abdullah Alsohim was doing an experiment at Reading examining this motility system and he needed some better photos for his PhD thesis. So he inoculated some surface spreading plates up with the double mutants on the Wednesday. He finished his experiment and obtained the photos on the Friday. But he forgot to throw his plates away, leaving them in the incubator. Over the course of the weekend, evolution worked its magic – the mutants were moving, and there looked like there were two variants on the same plate! The photo below shows the slow spreader emerged first with the fast spreader variant seen emerging on the left. Of course, Abdullah was perplexed because he thought he had done something wrong, but we realised that the mutants may have evolved motility – careful purification and re-testing the strains indeed confirmed this to be the case, where we found a slow moving strain emerged first followed by a fast mover. What we felt was remarkable was that FleQ was supposed to be a crucial chokepoint in the regulatory system – there was just no way the bacteria should be able to move again. But we know that in the face of adversity life can overcome these challenges.

Motility evolutionv1

To find out how this had happened we collaborated with David Studholme at Exeter University, who, in 2011, used his bioinformatics skills to analyse the re-sequenced genomes of the slow and fast moving mutants. He found the ntrB mutation in the slow mover and the same ntrB and subsequent ntrC mutation in the fast mover. When we did electron microscopy of the two strains, we were unable to see a flagellum for cells of the slow mover, whereas the fast mover exhibited an atypical “thick” flagellum. Closer examination of the latter strain appears to show multiple flagella intertwined together causing the thick flagellum structure near the cell and giving individual flagella filaments distal to the cell [see photo below].

Flg Evo ThickFlg evo EM

Next, by combining our microbiology know-how with the biochemistry and cell biology skills of Geraldine Mulley, and the evolutionary biology skills of Louise Johnson, Tiff Taylor and Mike Brockhurst we designed and carried out the crucial experiments needed to complete the story. Microarray data from Mike Brockhurst’s lab at York corroborated the motility observations, because we observed partial restoration of flagella gene expression in the slow moving mutant, and full restoration of flagella gene expression to wild-type levels in the fast-moving strain. The array results also showed a very strong up-regulation of genes involved in nitrogen uptake and metabolism in the slow mover, and whilst these genes were also up-regulated in the fast movers it was noticeably less than in the slow movers. This result was the key to understanding what was happening with the mutations in ntrB and ntrC. It has long been established in bacteriology that nitrogen gene expression is controlled by the NtrBC two-component pathway. When the cell needs more nitrogen, the NtrB protein transfers a phosphate group to NtrC, which acts in a similar way to FleQ by binding to enhancer sequences in DNA, recruiting RpoN and activating transcription; for NtrC these sequences are upstream of nitrogen genes. What we now believe is happening in the slow mover is that the ntrB mutation leads to a locked-on, overactive NtrB protein resulting in hyperphosphorylation of NtrC. Liam McGuffin found that this protein shares structural similarities with FleQ [see protein structural video post below] and we believe the level of phosphorylated NtrC becomes so high that it can start activating the flagellum genes (hotwiring the flagellum!) – however, it also activates all the nitrogen uptake and metabolism genes, causing a “toxicity” burden. That is when we see emergence of the fast spreader – an ntrC mutation occurs that we believe alters the DNA targeting specificity of the protein so that it begins to activate flagellum targets more than nitrogen targets. So the fast mover spreads faster on agar plates, either because more cells are making flagella or they are released from the nitrogen toxicity burden – or perhaps both.

Interestingly, we have since used light microscopy to examine cells of the mutants used in the study. The double mutant strain cannot move, the slow mover seems to be a “mixed” population of flagellum producers and non-producers, while the fast movers were swimming around at high speed. Intriguingly, the fast mover also showed a clumping phenotype, although the clumps appeared to be moving. This may be occurring because the cells make multiple flagella and get tangled together.

In parallel to this, our collaboration with Mark Silby and Alex Dills at University of Massachusetts Dartmouth was yielding exciting results, especially when they discovered two alternative mutational routes leading to over activation of NtrC. This was a pivotal moment in the project because it showed that the evolutionary rewiring was rapid and repeatable in two different systems, but could also follow slightly different trajectories in the slow mover, converging on the subsequent mutation of NtrC in the fast mover.

In summary, the bacteria have effectively evolved the ability to swim by hotwiring the nitrogen regulatory system to restart their faulty motility system. To find this out we had to work across traditional disciplines, bridging microbiology, evolution and biochemistry, to build a truly interdisciplinary team. Of course, there are many gaps in our knowledge, one of which is the likelihood of these changes occurring in a real life situation. For Pseudomonas fluorescens real life involves swimming in soil water and moving across plant roots, not agar plates. Nevertheless, we believe our discovery is an amazing tale about life’s resilience, that in the face of catastrophic mutation threatening bacterial extinction, there is still the capacity for innovative changes to rescue the mutants and save the day.

Finally, some thanks to people who have kindly provided encouraging feedback along the way: Prof. Martin Buck, Prof. Allan Downie, Prof. Paul Rainey, Prof. Jim Alfano, Prof. John Mansfield, Prof. Jeff Dangl, Prof. Joyce Loper, Dr Gail Preston, Dr Joerg Schumacher, Dr Xue-Xian Zhang, Prof. Ben Whalley, Prof. Ian Jones.

The route to rewiring – lay background

A popular quip is attributed to Otto von Bismarck: if you like laws or sausages, never watch them being made. Scientific papers, we’d argue, are more like sushi rolls than sausages: you can sometimes appreciate them even more if you see the work taken to put them together. We’re also really keen on communicating science to everyone [ not just to our colleagues. So, following the release of our paper in the current issue of Science [], we’re providing some explanation and backstory via this post.

Microbiologist Rob Jackson has worked on Pseudomonas fluorescens for over 10 years. They’re common bacteria in soil, where they help plants grow and protect them from disease – so they may one day be able to help us grow crops in a more efficient, environmentally friendly way. Together with his then PhD student, Abdullah Alsohim, Rob was trying to understand how P. fluorescens moves around in soil and reaches plant roots. They knew that it could swim using flagella – complex, multifunctional structures that many bacteria use to swim, anchor themselves or sense their environment.

In order to compare the behaviour of bacteria with and without flagella, Rob and Abdullah used a genetically-engineered, flagella-less strain of P. fluorescens. This strain was missing a gene called fleQ, which is a control gene: it acts as a switch, kick-starting all the other genes for parts and assembly of the flagella. Abdullah discovered that the genetically engineered mutants could still move, although differently: on an agar plate, they formed colonies with distinctive tendrils. It turned out that they had a second, backup way to move around, by oozing a lubricant called viscosin and sliding about on it. Abdullah then removed the gene for viscosin as well, creating new, totally immotile bacteria. Placed on an agar plate, his new strain, called AR2, would remain stuck in the middle, unable to spread out to the edges.

Next came a key step in our research: we made a chance discovery. Rather than growing his AR2 bacteria overnight as usual, Abdullah accidentally left some plates growing over a long weekend. By Monday, the bacteria were moving again. Abdullah thought at first that he had done something wrong, but we realised that we could be seeing rapid evolution, and when we repeated the experiment (intentionally this time) the same thing happened – the bacteria started to move, first slowly and then fast. Under an electron microscope, the fast swimmer had clearly regained its flagellum.

Luckily enough, around this time the Jackson lab started a new collaboration with evolutionary biologists. Louise Johnson had recently won a Leverhulme grant, along with Rob and with Mike Brockhurst from the University of York, and they had hired Tiffany Taylor for a separate project involving microbiology and real-time observation of evolution in action in the lab. We had the right team together to recognise and address interesting questions in evolutionary biology – so when Abdullah completed his PhD and moved away, Tiffany was the perfect choice to take over the reins and take the project to a new level by asking how this happened. With the FleQ switch removed, it should have been completely impossible for the flagellum to reappear, let alone so quickly.

To find out how the bacteria had regained flagella, we went straight to their genes. Our collaborator David Studholme at Exeter University analysed the slow- and fast- swimming strains, and found mutations in two new genes: both of them had a mutant NtrB gene, while the fast swimmer had a second mutation in the NtrC gene. These two genes work as a pair to control nitrogen intake: the cell can sense when it’s starving using NtrB, and suck up more nutrients if necessary using NtrC. Like FleQ, NtrC is a control gene, but it normally switches on feeding genes rather than flagella genes.

Because mutation is a random process, it’s often been thought that evolution will tend to work in an unpredictable, meandering way. But when we re-ran the experiment several times, the same genes mutated – sometimes in exactly the same places.

Our next stroke of luck was having Geraldine Mulley in the lab next door. Geraldine had worked on this kind of paired genes before, and she helped us design experiments to test our fast and slow swimmers with different food sources, to see what the two Ntr mutations did. Talking over our results with her, we sketched out our thoughts about what was going on inside our mutants, and then to make sure we’d got the right idea, we measured the activity of all the genes in each mutant with the help of Mike’s colleague Peter Ashton in York.


This scribble from Louise Johnson’s notebook shows how we came to understand the re-evolution. The left hand scribble is the immotile ancestor. Its Ntr genes are doing their usual job, but with fleQ removed, nothing can switch on the flagella genes. In the middle is the slow swimmer: regardless of actual food availability, the mutant NtrB gives out a strong I’m starving signal, resulting in an overload of NtrC. This gives the cell a good hard kick in the biochemistry: it switches on the feeding genes so aggressively that these cells will “overeat” and poison themselves, but it also spills over slightly to trigger the flagella genes. On the right, the fast swimmer’s mutant NtrC has adapted to its new job, and become a multitasker: it doesn’t overactivate the feeding genes so much, but efficiently switches on the flagella.

When Mark Silby and Alex Dills from the University of Massachusetts Dartmouth repeated the experiments in a different variety of P. fluorescens, they found many of the same mutations as we did, but also discovered an alternative first step to the evolutionary pathway: sometimes the first mutation was to one of the feeding genes themselves, so that the cell is actually starving.

Liam McGuffin, a protein bioinformatician, completed the team. His computer programs worked out how the DNA mutations changed the shape of proteins, and showed us the remarkable similarity in shape between the switch that was lost, fleQ, and the switch that was substituted, ntrC – the two are distant molecular cousins [see protein structure video below]. His work joined the dots so that we could trace the process all the way from the DNA, through the proteins and biochemical interactions, to the cell’s behaviour and its evolutionary advantage. We think this is a really nice example of evolution in action: each step is a kludge, and a kludge can bring serious drawbacks – but small steps soon add up to a result that seemed impossible.

Weekend evolution: Bacteria ‘rewire their genes’ to fix a faulty motor

An immotile bacterium (left) evolves in just a few days until it regains some (centre) and then full mobility (right), including the use of a tail-like flagellum

An immotile bacterium (left) evolves in just a few days until it regains some (centre) and then full mobility (right), including the use of a tail-like flagellum

Scientists have found how simple bacteria can restart their ‘outboard motor’ by hotwiring their own genes.

Unable to move and facing starvation, the bacteria evolve a replacement flagellum – a rotating tail-like structure which acts like an outboard motor – by patching together a new genetic switch with borrowed parts.

The findings, published in the journal Science, show that when an organism suffers a life-threatening mutation, it can rapidly rewire its genes. The remarkable speed with which old genes take on new tasks suggests that life has unexpected levels of genetic flexibility.

The discovery was made by a team at the University of Reading, led by Dr Robert Jackson and Dr Louise Johnson. The Reading team, funded in part by BBSRC and the Leverhulme Trust, collaborated with scientists at the universities of York, Exeter, and Massachusetts Dartmouth.

Dr Tiffany Taylor, University of Reading, joint lead author of the study, said: “Our findings in this study show how resilient life can be.

“We thought we had broken the bacteria’s ability to move beyond repair. In theory, the bacteria should have starved to death and effectively gone extinct. Yet over the course of a weekend they managed to patch themselves back together with borrowed genes.”

Happy accident

Scientists made the discovery by accident, while researching ways to use naturally occurring bacteria to improve the yield of crops. A microbe was engineered so that it could not make its ‘propeller-like’ flagellum and forage for food. However, when a researcher accidentally left the immotile strain out on a lab bench, the team discovered the bacteria had evolved over just a few days. The new variety of bacteria had resurrected their flagella in the process.

Remarkably, this happened because the mutants had rewired a cellular switch, which normally controls nitrogen levels in the cell, to activate the flagellum. This rescued these bacteria, which faced certain death if they didn’t move to new food sources.

The bacteria being studied, Pseudomonas fluorescens, are among a group of bacteria scientists are researching for use in agriculture, as a kind of ‘plant probiotic’. These could help crops grow or fight off diseases, leading to higher yields. However, a key problem is that the bacteria lack resilience, as their positive effects can stop working after only a short period of time.

Dr Jackson, a microbiologist at Reading, said: “Plant probiotics could make crops grow more reliably in the future, helping to feed the world’s growing population. This new study shows that these bacteria are more resilient than previously thought, as they show a remarkable capacity to overcome catastrophic changes and find a way to survive.

“This gives us crucial insights into how bacteria could survive and change, and the challenge now is to see if this occurs in their natural soil and plant environment.”

Borrowed parts

Dr Johnson, an evolutionary biologist at Reading, said: “Evolution has been described as a process of ‘tinkering’, but this work shows that evolution can be remarkably repeatable. When the situation is desperate, life finds a way.”

Dr Liam McGuffin, a University of Reading bioinformatician, built three-dimensional computer models of the proteins involved, which showed that the missing molecule and its new substitute were very similar.

“The bacteria effectively finds a similar protein molecule to the one we knocked out and uses it for a different purpose,” Dr Taylor said. “This molecule then gets used as a replacement ‘key’ to ‘hotwire’ its motor.

“But the hotwiring comes at a cost. The replacement key is a molecule borrowed from a system which regulates nitrogen levels. The mutant bacteria can now move, but it can’t regulate nitrogen properly, which can build up and become toxic. Of course, it’s an evolutionary price worth paying when the alternative is certain death.”

This research will appear in the 27 February, 2015, issue of the journal Science, published by the AAAS, the science society, the world’s largest general scientific organization. See, and also

Full reference: Taylor T. B., Mulley, G., Dills A. H., Alsohim A. S., McGuffin L. J., Studholme D. J., Silby M. W., Brockhurst M. A., JohnsonL. J., Jackson R. W. ‘Evolutionary resurrection of flagellar motility via rewiring of the nitrogen regulation system‘. Science, 347:1014-17, 2015