- Wallden & Fange et al. Cell, 2016. The Synchronization of Replication and Division Cycles in Individual E. coli Cells.
- Hammar & Wallden et al. Nature Genetics, 2014. Direct measurement of transcription factor dissociation excludes a simple operator occupancy model for gene regulation
- Persson et al. Nature Methods, 2013. Extracting intracellular diffusive states and transition rates from single molecule tracking data.
- Fange et al. Nature Methods, 2012. Lost in presumption: stochastic reactions in spatial models.
- Hammar et al. Science, 2012. The lac repressor displays facilitated diffusion in living cells.
For more information & complete list of publications, please visit the library.
The overall ambition of our research is to bridge the gap between quantitative physical models and biological observations in order to identify and resolve inconsistencies in our current understanding of life at the molecular level.
We are particularly interested in how key steps in transcription, translation and replication are regulated in the intracellular environment and at what level of physical detail these processes need to be modeled to describe their function in the living cell.
To answer these questions we use state-of-the-art single molecule microscopy methods to study kinetics and diffusion in living cells (Elf et al Science 2007, 2012, PNAS 2011, Hammar et al. Science 2012). We also develop new microfluidic techniques and corresponding analysis software to control growth conditions and enable automation. These experimental techniques are accompanied by the development of pioneering computational tools for stochastic reaction-diffusion simulation of intracellular kinetics (Fange and Elf, PLoS CB 2006, PNAS 2010, Fange et al. Nature Methods 2012) and mathematical modeling of intercellular physiology (Elf et al, Science 2003; Nature Physics 2009, PNAS 2010, Nature Communications 2011).
Johan Elf has received a VR grant of 24 million SEK over 6 years to build a research environment focusing on the physical principles for genome search by means of large scale dynamic phenotyping of strain libraries. The environment also comprises the Tenje group at the Department of Engineering Sciences, who works on microfluidic solutions for cell analysis, the Kamerlin group at the Department of Cell and Molecular biology who performs MD simulations of protein-DNA interactions and the Johansson and Deindl groups at the Department of Cell and Molecular biology that are both investigating single molecule interactions and kinetics using fluorescent microscopy.
The interdisciplinary environment will build bridges between biology, engineering and physics with many potential applications throughout the life sciences. Appealing as it may sound, collaborative efforts between engineers and scientists are rare in Sweden and much is to be gained by closing the loop between technology development and fundamental research.
If successful, the technology that we are developing will make it possible to connect dynamic phenotypes to the underlying genotype for thousands of different cell strains simultaneously. This will have important applications in many areas of science including drug discovery, biotech product optimization and basic genetic research.
By following single molecules diffusing in the cell we can learn much about how biomolecules interact with each other in the cytoplasm. How the molecule switch between diffusive states can be used for example to estimate the reaction rate constants for molecules in equilibrium. One problem, however, has been that irreversible photo-bleaching greatly limits the individual trajectory length and thus the dynamic range of the reactions that can be studied. Historically, the only way to increase the trajectory length has been to reduce the light intensity and not collect so many photons per image, but as localization accuracy correlate negatively with the number of photons it has always been at the expense of the spatial resolution.
In collaboration with Stefan Hell's group in Goettingen we have fundamentally changed the prerequisites for single molecule tracking in living cells. Instead of imaging the light from a molecule and then use the point spread function to appreciate its spatial location, we now follow the movement of the molecule using a hollow laser beam. The fluorescent molecule is trapped in the center of the laser beam where it is not excited and thus not emitting any photons. Only when the fluorophore deviates from the middle, we detect photons that we can then use to adjust the position of the laser and bring the molecule back to the center. Since the position of the molecule is defined indirectly by the laser position, we can use the available photons much more efficiently. In principle, one detected photon is enough to know that we are out of place, and only a few more are needed to calculate a new optimal position. This means that we can record trajectory lengths that are 100 times longer than what have been possible using standard techniques. The downside is that the system is complicated to set up since laser correction must occur on the microsecond scale.
The new MINFLUX microscope will make it possible to finally study biological processes on the same spatial and temporal scale that can be reached by molecular dynamics simulations. The system that we have built in Uppsala will initially be used to study gene regulation and protein synthesis in living cells. An exciting journey into living matter at the molecular level has just begun!
Read the complete article in Science.
Scientific conferences are serious bussiness. No explanation necessary.
The progress in the field of live cell single molecule fluorescence microscopy that has taken place over the last couple of decades has been spectacular. However, good protocols for labeling the molecule of interest with a small, bright and photo stable fluorescent dye is limiting progress in the field, and the labeling step is where most live cell experiments fail.
We have now explored ways to use of noncannonical amino acids to introduce handles where small organic fluorophores can be bound specifically to the protein of interest. The noncannonical amino acid is encoded via the pyrrolysyl-tRNA/pyrrolysyl-RNA synthetase pair at artificially introduced TAG codons in E. coli strains that have been recoded to lack endogenous TAG codons as well as the TAG-specific release factor RF1. The amino acids contain bioorthogonal groups that can be clicked to externally supplied dyes, thus enabling protein-specific labeling in live cells with small bright fluorophores.
The incorporation of the noncannonical amino acid in the protein of interest works well. The overall labeling scheme works well works well for proteins that spend some time on the surface of the cell and thus can be clicked to the dye extracellularly. As a model protein for the NcAA-based small fluorophore labeling we use the outer membrane porin OmpC, one of the most abundant outer membrane proteins in E.coli, which is involved in the control of cellular osmolarity and the uptake of nutrients and antibiotics. The osmoporin OmpC can be labeled sufficiently specific to enable single molecule particle tracking.
In order to label proteins in the cytoplasm however, the dye has to cross the cell membrane, which proved to be surprisingly difficult. Hydrophobic dyes can cross the membrane relatively easy, but they also interact nonspecifically with various elements in the cell and thus cause high background levels. Hydrophilic dyes are less sticky but must be forced across the membrane, a process that has severe consequences on cell viability.
Read the complete article in ACS Synthetic Biology.
A double strand break in the chromosome means trouble for any organism and the only hope resides in finding an intact copy of the break side on another chromosome. When this homologues sequence is localized, a genetic exchange takes place and both chromosomes walk away with one complete sequence. The cell is saved.
Science has presented detailed descriptions of the different steps in this homologous recombination process, all but one. According to the classical model, the broken ends randomly probe the DNA by base paring in search for the homologous sequence. If the correct site is not in absolute proximity, weeks of interrogation would be required even with a chromosome diffusing like a small molecule. Needless to say, it is not.
If the classical model is correct, homologues recombination would simply not work when the homologues sequences are not close, yet an overwhelming amount of experimental data claims that it does. We can only speculate at what might resolve this inconsistency, but a parallelized search by DNA or RNA scouts somehow produced at the break site is an attractive hypothesis. A thorough examination of available experimental observation reveals details that make this solution all the more likely. However, experimental verification of the theory is yet a challenge... Any takers?
A complete description of the hypothesis is presented in Cell systems.
Have you ever wondered why DNA is a helical molecule? Although the shape of this ultimate information carrier is maybe an inevitable consequence of stacking single monomer units into a polymer, the shape also has an important effect on the time it takes for DNA binding proteins to find their targets. By solving the reaction-diffusion equations for DNA-like geometries, and complementing with simulations when necessary, we show that the helical structure can make binding to the DNA more than twice as fast compared to binding a straight molecule.
Figure: Reaction-diffusion equation and boundary conditions for example geometry. Only when the reactive patch on the sphere (gray) is in contact with the reactive patch on the cylinder do they react with rate kappa. The sphere moves around the cylinder with translational diffusion rate D. When the sphere reaches a distance Rc away from the cylinder it is equally likely to bind another DNA segment.
Check out the publication in Journal of Physics A.
How the cell cycle is regulated in bacteria has been examined and debated since the first model was put forth by Cooper and Helmstetter in the 1960s. Using a combination of microfluidics and image analysis of E. coli with fluorescent labels on the replication machinery, we have been able to study how the replication and division cycles are coupled in individual cells. This made it possible sort out what is the source if variability in cell sizes and generation times. In brief, all cells initiate at a constant volume per chromosome. The time from initiation to division is given by the individual cells growth rate which is variable and is the major cause of the variation in cell sizes. It has been a long project, 5 years, but it turned out very nicely in the end.
(Top) The cell volume (black solid line) of a single cell lineage simulated according to the single-cell CH model throughout an upshift in growth conditions. The growth rates are sampled from the observed distributions. When the cell passes the initiation volume(s) (red lines), a division event is triggered (green arrow) after a time period corresponding to the C and D periods (gray bars). Variation can be introduced in growth rates, initiation volume, or C+D period. (Bottom) Demonstration of how the model performance is improved as cell-to-cell variation is introduced in growth rate (mu) as well as in initiation volume (V) and C+D period (tao).
Check out the publication in Cell.
At the General Meeting of KVA on May 11, Johan Elf was elected member of the academy's chemistry class.
Read more about the Royal Swedish Academy of Science.
But do not worry! The latest addition to the Elfware family will not hiss in your ears when you sleep. Not unless you have a bias in your image analysis, that is ;)
SMeagol will help you optimize your experimental set-ups and make sure that there is no bias in your data analysis. The software creates realistic synthetic microscopy images by combining reaction diffusion simulations with simulations of the fluorophore photophyscis and the photophysics of the optical system. SMeagol will be presented in more detail in an upcoming issue of Bioinformatics.
(Left) Bright field image of a bacterial cell with simulated "true" locations of the molecules to be imaged and tracked. (Middle) Experimental background noise. (Right) Simulation of how the data might look like given the experimental parameters.
It was a few minutes to one and I would be lying if I said that the atmosphere in C8:301 was relaxed. The past few weeks had seen a desperate activity when more than four years of pioneering scientific work was to be summarized, put in context and, unfortunately, explained to the less talented. If it had only been a question of unifying the theory of micro and macroscopic interactions of proteins and DNA in a scientifically appealing manner, Anel Mahmutovic would have been home free. However, science, it had lately become painfully clear, was also about fonts, spelling and dinner arrangements…. Who would have guessed? The opponent, professor ten Woldemort, adjusted his velvet hat and made sure that his notes were in order. Finally, the clock struck one and the battle begun. G In an instance, the PhD student who trebled in the face of the fearsome professor ten Woldemort was gone, and rising from the ashes was a collected and well-spoken young scientist who met the gaze of his opponent and answered his question without the slightest tremor in his voice. A few hours later, his friends, family and colleagues congratulated Dr. Anel Mahmutovic to a job well done and a title well earned. In the evening, at first hotel Linné, it was a party wanting nothing and Anel was in every sense the perfect host he swore he’d never be.
We have previously measured the sliding distance of LacI on DNA in vivo and concluded that the transcription factor uses a combination of 3D diffusion in the cytoplasm and 1D diffusion on the DNA to speed up the search for the operator. We now dig deeper into the sliding mechanism to figure out more about how the TF slides on DNA. How fast is the sliding? Is it smooth or does it include frequent microscopic jumps (hopping)? What is the influence of multiple DNA binding domains allowing the TF to transfer between DNA strands and what is the effect of molecular crowding? Using Monte Carlo Simulations scheme to realize the search process for one or two operator sites, we try to define a physically reasonable sets of parameters for which the model can explain the observed in vivo measurements.
We find that non-specific binding to DNA is improbable at first contact and that the sliding LacI protein binds at high probability when reaching the operator. We also conclude that the contribution of hopping to the overall search speed is negligible although physically unavoidable and we see an unexpectedly high 1D diffusion constant on non-specific DNA sequences. Including a mechanism of inter-segment transfer between distant DNA segments does not bring down the 1D diffusion to the expected fraction of the in vitro value, suggesting a more complicated explanation than molecular crowding.
Look out for the full article by Anel Mahmutovic that will soon be available in Nucleic Acid Research
Elflab Ski conference 2015 offered everything you could wish for in terms of scientific discussion, excellent international food, snow blizzard skiing and ambulance drivers with no sense of direction. The only thing missing was Mats, Daniel and Elias. And Cia, who spent the week-end giving birth. Big congratulations to a beautiful baby boy!
Beautiful chistmasy e. coli captured by Prune Leroy.
In the "Nobel studio" second episode, Swedish television presents personal portraits of the Nobel laureates in chemistry 2014 and also tries their best to explain the wonders of single molecule microscopy. We also get to share Johan Elf's view on football and we learn what popular faces might actually hide in our morning toast. But hurry, the program is only available until the 9th of January, 2015.
The Nobel Prize in chemistry 2014 has brought a lot of attention to the application of super resolution microscopy. Read what Åke Spross write about the research in the Elflab in UNT's Sunday edition.
How precise is cyclic life? That was the question Mats Walldén set out to answer on this cloudy December morning. After a brilliant performance from both Mats and his opponent, prof. Peter Graumann, the answer was presented; life is 20% precise. With reference to another fine piece of literature, an ultimate answer is not always as satisfying as you might think. For a more complete understanding you might read the thesis in its entirety, however, if you cannot seem to find the time, I recommend the first section life (page 11-12) and the conclusions and future outlooks (page 60-63). Below are a few lines from the introductory paragraph, a rather dystopic view of the unicellular situation. You cannot help but wondering; is there nothing we can do for these poor creatures?
"Gravity is all but suspended and instead of falling to the earth a small object is continually bombarded by its environment, forcing it to perform a random walk, to diffuse, if not attached or actively propelling itself. Inertia is negligible, so that movement stops immediately when active propulsion is discontinued. Ballistic weaponry is exchanged for a chemical arsenal as collisions carry no momentum. Sexual reproduction is rare and caring for progeny is rarer still."
What is the importance of super resolved fluorescence microscopy for the research in the Elflab and for our society in general, what are the life-stories of the Nobel laureates and what is the next challenge for life science microscopy? Listen to Johan Elf being interviewed by the journal Expressen.
We almost take it for granted, the fact that we can peer into the nano-world and study the life and beauty of individual molecules as they go about their business in the crowded interior of the biological cell. Not so long ago this was impossible... theoretically. The theory was not wrong, but luckily there were people that refused to be held back by the seemingly impossible and eventually found a way around the diffraction limit for spatial resolution. Some of these brave scientists were awarded the Nobel Prize in Chemistry for their efforts. Some were not. We are grateful to all of you!Nobel Prize in Chemistry 2014.
After a brilliant introduction of the field by opponent Achillefs Kapanidis (Oxford) we enjoyed a scientific battle between him and Arash that lasted a good two hours (top). Afterwards the spirit remained high as Arash and co awaited the decision of the committee (bottom).
The thesis describes single-particle tracking, a technique that allows for quantitative analysis of the localization and movement of particles. Recent advances have made it possible to track hundreds of particles in an individual cell by labeling the particles of interest with photoactivatable or photoconvertible fluorescent proteins and tracking one or a few at a time. Arash show that the fluorescent protein mEos2 diffuses normally at 13 µm2/s in the E. coli cytoplasm and also that free ribosomal subunits have access to the bacterial nucleoid where fully assembled ribosomes are excluded. The thesis "Biological Insights from Single-Particle Tracking in Living Cells" can be read here.
In this work, that was recently published in PNAS, we track fluorescently labeled ribosomal subunits in bacterial cells. We show that translating ribosomes move much slower than free subunits. More importantly, we show that it is only translating ribosomes that are excluded from the bacterial nucleoid, whereas free subunits have full access to the nucleoid. This finding is important because several gene-regulation mechanisms require that ribosomes are able to initiate translation as soon as the RNA polymerase has initiated transcription, and it has been difficult to reconcile such a requirement with the observation that ribosomes are nucleoid-excluded.
The intensity maps show the location of bound and free ribosomal subunits. Although the fully assembled ribosome is excluded from the nucleus, the small subunits have access and can initiate translation before the complex is translocated to the periphery.
The Elf lab enjoys an outing on the river Fyris. From the top, left to right: Arvid, Prune, Tom, Kalle, David, Jimmy, Ebba, Cia, Fredrik, Ozzy, Johan, Frederik, Irmeli and Mike.
Read what Valda Vinson write about Petter and Mats' paper "Non-equilibrium contributions to transcription factor mediated gene regulation" in Science Magazine editors' choice, Volume 343, Number 6178, Issue of 28 March 2014.
A new method for direct measurement of transcription factor dissociation makes it possible to exclude a simple operator occupancy model for gene regulation.
One by one the mysteries of gene regulation are subject to scrutiny, and our models, the way we understand gene regulation, are put to the test. We know that lacI, the repressor of the lactose digestion machinery in the cell, searches for its binding site using a combination of 3D diffusion and sliding on the DNA, we know that it slides around 45 bases before detaching and we know that it takes less than a minute for a lacI repressor to find and bind the operator, although most of the time, the TFs slide over the binding site without binding. However, to test the prevailing model for TF mediated gene regulation, which assumes that the level of repression is determined by the equilibrium binding by the repressor to its operator, we also need to know how long the TF stays bound to the DNA before it dissociates macroscopically, e.i. leaves the immediate vicinity of the regulatory site. Using a new technique based on a single molecule chase assay we can answer this question - lacI stays bound to its specific operator for on average 5 min. So far so good, with an association time of 30 s and a repression ratio of 10 (as measured by an enzymatic reporter) the model is in no immediate danger. However, if the native operator is replaced by a stronger one, this results in a slower dissociation while the association rate stays approximately the same. Conclusion; our findings do not support the simple equilibrium model and this discrepancy has to be considered when predicting gene activity from TF binding strengths. For more details, visit Nature genetics, where the study can be read in its entirety.
Petter Hammar is awarded Bjurzons Premium for his thesis lac of time. The prize is awarded annually by the Uppsala University Vice-Chancellor in recognition of an excellent scientific thesis by a student or young lecturer.
What do elephants have to do with anything? Find out more in Petter's thesis
The Christmas spirit is particularly strong in the laser-lab where the new 638 nm laser takes up the competition with Rudolph the reindeer in Christmas-red glowiness.
We have previously shown that the lac repressor slides along the DNA to speed up the search for its specific binding site. By combining micro- and macroscopic computational models, we have now been able to investigate the details of the sliding mechanism on the atomic scale. The lac dimer follows the major groove of the DNA helix, causing it to slide in a spiral motion. It remains close to the DNA for about 8bp before making a microscopic dissociation and stays on the same DNA fragment for an average 48 ms. This corresponds to an in vitro sliding length of 240 bp which correlates well with in vitro measurements. The study presents an unique combination of theoretic tools and the results satisfyingly connect macro-/mesoscopic events at the nanometer level, which enables a deeper interpretation of experimental observations.
On a rather rainy day in June, Petter Hammar and his opponent Antoine van Oijen put on a show that kept the audience enchanted for a good 3 hours. The celebration continued at the Old Fellow House well into the following morning.
Left: Petter and Antoine deep in discussion. Top right: The committee delivers the happy news that Petter has passed the examination. From the right, Antoine van Oijen (opponent), Carolina Wählby, Yu Ji, Nora Ausmees and Johan Elf (proud supervisor). Paul Blainey and Mats Nilsson were also part of the evaluation committee. Bottom Right: The defense attracted a rather large crowd.
Welcome to an afternoon with excellent talks on the topic of new techniques to study the fine details of life at the molecular level:
It is easy to imagine that the transcription factors (TFs) are superfast regulators of gene expression, but in reality it might take several minutes before a TF finds and binds its target sequence. As a result, negative feed-back (where the expression of a TF is inhibited by the TF itself) cannot at the same time be fast and strong.
If the TF binds strongly to the repressor site, most proteins are necessarily produced immediately after cell replication and if the binding is very week, the protein production is essentially linear; in both cases the negative-feedback is virtually nonexistent. In a recent paper in Nature Communications, we show that there is an optimal TF binding strength, where the time the binding site is free is not dependent on TF concentration. An important implication is that the TF binding strength is not necessarily correlated to its functional importance as a gene regulator.
Single molecule tracking data usually contains large amounts of information, but extracting the data from the highly fragmented trajectories, which is often the result of SPT in vivo, can be a real challenge. Traditional methods use Mean Squared Displacement and/or Cumulative Distribution Function analysis to identify parameters in presumed underlying models, but the model has to be guessed a priori and introduction of additional states does always lead to a better fit.
In a paper that was recently published in Nature Methods, we describe an analytical tool based on a variational Bayesian treatment of hidden Markov models that combines the information from thousands of short single-molecule trajectories of intracellularly diffusing proteins. The method identifies the diffusion constants and state transition rates as well as the number of states in the model.
Using this method we have created an objective interaction map for Hfq, a protein that mediates interactions between small regulatory RNAs (green) and their mRNA targets (grey), see image to the right. Photoconvertable proteins were used to track single hfq molecules (yellow) and assign them to different kinetic states based on their diffusive properties. The diffusion constant of hfq depends on its state of binding. Free hfq diffuse fast, but when the molecule is bond to other molecules, e.g. mRNA, the molecule is slowed down. The image was featured on the cover of the March issue of Nature Methods.
After a semester of hard endeavors the Lab Elfs take off to get some well-deserved rest...
...or will they?
Johan Elf is appointed Professor in Biological Physics at Uppsala University.
During recent years, physical modeling has become increasingly important to generate insights into intracellular processes. Many times it is essential to consider both the spatial and stochastic nature of chemical reactions to be able to capture the relevant dynamics of biochemical systems. In this review, which is currently in press for Nature Methods , Fange and Mahmutovic discuss when to use, and when not to use, different models to achieve the best possible balance between speed and physical accuracy.
Different levels of quantitative modeling frameworks for intracellular chemistry.
MesoRD is a simulation tool developed in the Elf Lab. The software is used to simulate stochastic reaction-diffusion systems as modeled by the reaction diffusion master equation. The simulated systems are defined in the Systems Biology Markup Language with additions to define compartment geometries.
Apart from bugfixing, the new release of MesoRD updates the code for the microscopically treated bi-molecular reactions to also include reactants with different diffusion rates.Download MesoRD-1.1 here
Gustaf and Mats' paper "Hi-throughput gene expression analysis at the level of single proteins using a microfluidic turbidostat and automated cell tracking" is published in Philosophical Transactions B.
In order to confidently draw conclusions on the nature of transcriptional diversity it is necessary to sample a large number of cells. If done manually this could easily amount to weeks of analysis. By combining automated cell tracking with a microfluidic culture chamber this method makes it possible to analyze the rate of gene expression at the level of single proteins in a sufficient number of bacterial cells.
The movie to the right shows automated tracking of E. Coli in the microfluidic turbidostat
The lac repressor is a protein that binds to specific DNA sequences on the E. coli choromsome and regulates the activity of genes. In order to rapidly find these DNA sequenes among millions of others, we have shown that the lac repressor combines siding along the DNA sequences and free diffusion in the cytoplasm. See illustration by Tremani/Elf
Link to abstract at Science website.
Top row from left: Mats, Sorin, Anel, Petter, Fredrik, David and Johan.
Front row from left: Arash, Gustaf, Prune and Cia. Missing: Erik and Andreas.
The Elf lab got the highly competitive ERC Starting Independent Research Grant. Less than 3% of the 9200 applications were granted. The grant implies that a number of postdocs and PhD students will be recruited over the next few years.
We have started building our first single molecule microscope for live cell imaging
J Elf was awarded the Ingvar Carlsson Award (3MSEK) by the Swedish foundation for strategic reserach.
Dr. Brian English joins the Elflab as a postdoctoral fellow. Dr. English is an expert in single molecule sepectroscopy and got his PhD in chemistry at Harvard University.
Returning from the Xie group at Harvard I am now starting my own group in the department of Cell an Molecular Biology at Uppsala University.
Our work will be focused at the development of new experimental and computational methods for analyzing intracellular transcription factor dynamics at high temporal resolution and spatial precision.
The methods will be used to answer fundamental questions in bacterial physiology related to transcription factor mediated gene regulation in living bacterial cells.