To manipulate their host or ensure survival within, many bacterial pathogens export molecular toxins  via so-called secretion systems. The type III secretion system (T3SS), also known as injectisome, resembles a molecular syringe. It is used by Salmonella, Shigella, Yersinia and other pathogens to inject effector proteins directly into human cells. The T3SS consists of a hollow needle, a basal body spanning both bacterial membranes and a set of conserved cytosolic components at the interface of the inner membrane. Since the first injectisome was visualized 25 years ago, we have learned a lot about the mode of action, structure, and assembly pathway of the T3SS. However, the central question of how effector proteins are recruited to the machinery was way less understood.

Prime candidates for this task are the cytosolic T3SS components, which were shown to act as a sorting platform in a landmark publication in 2011. Our lab had recently discovered that the cytosolic components are present in two states – while some proteins are bound to the injectisome, the larger part is mobile in the cytosol. Strikingly, both pools exchange and this exchange correlates with the secretion of effectors. But could these dynamic cytosolic components act as a kind of effector shuttle, picking their cargo up in the cytosol and faithfully delivering it to the export apparatus? And how could we test if and where the effectors directly interact with the cytosolic components? Past co-purification attempts had only yielded limited insight, possibly because of the transient interactions characteristic for transport processes: The interactions between the cytosolic injectisome components and their effectors need to be stable enough for shuttling effectors from the cytosol to the membrane-bound needle complex, but also have to allow for the efficient release of the effector substrates for export. This makes these complexes very short-lived and difficult to detect in standard in vitro biochemistry techniques that rely on purified components.

One day back in 2017, at the start of my PhD studies, my supervisor Andreas Diepold, a freshly established PI at the Max Planck Institute in Marburg, came from a meeting with the neighboring lab of Ulrike Endesfelder. He proposed to track the cytosolic components with single molecule localization microscopy (SMLM) in presence and absence of effectors in Yersinia enterocolitica, a relative of the plague bacterium and the main model organism in our lab. If these proteins interact in the cytosol, the presence of effectors should slow down their diffusion speed.

After identifying suitable culturing and washing conditions for SMLM (a technique that – due to its sensitive single-molecule detection ability – is very sensitive to any fluorescent contaminants left in the sample), everything was set up for a first test. Bartosz Turkowyd, a PhD student from the Endesfelder lab, was operating the microscope and we were very excited to see the T3SS components, with clusters at the membrane (being bound to the needle complex) and various mobile signals in the cytosol (corresponding to complexes made from different numbers of T3SS components with and without effectors as we now know). Although I have seen this many times since then, the initial moments of a new sample visualizing the behavior of single molecules live on the microscope never fail to amaze me.

Since the first results seemed promising, increasing the throughput of the experiments was next. Alexander Balinovic, who had just started his PhD in the Endesfelder lab, took over the imaging and analysis responsibility. Long incubation times, temperature and media switches and the added challenge that after preparation, the samples needed to be sealed and carried from our lab to the custom-made SMLM microscope in a neighboring building, meant that measurement days were very long and required lots of planning. Without the support of two dedicated Master students in our lab, Carlos Helbig and Moritz Fleck, it would have been impossible to measure these many samples in such a short time.

We had already noted during the first experiments that indeed, the presence of effector proteins slows down the diffusion of the cytosolic components. To analyze these interactions in a more controlled, but still biologically relevant setting, we then performed experiments in bacteria lacking any of the T3SS components, systematically introducing specific pairs of potential interacting partners. Calculating the diffusion speeds of combinations of cytosolic components, based on their molecular mass and the Stokes Einstein equation allowed us to predict different oligomeric states and combinations of cytosolic subunits. Using those, we were able to correlate the observed diffusion measurements to different oligomeric states and complexes made of different combinations of components, which cleared up the picture of what we were looking at. How well this correlation held despite all the factors influencing diffusion in living cells remains astounding.

However, our collaboration was about to become far more difficult: After three years of working together on this project, in 2020, the Endesfelder lab moved to Carnegie Mellon University in Pittsburgh while the Covid-19 pandemic hit. Despite our best efforts, measurements were not completed when our collaborators and friends left for their new home. Although we had obtained funding by the German Academic Exchange Service to finish the measurements in the US, boarders were sealed, traveling rules constantly changed, and the American embassy in Frankfurt was all but closed down to the public. Getting all the paper work done for a J-1 visa under those circumstances turned out to be challenging and took a lot of effort, time, and nerves (involving, e.g., rechecking the scheduling website of the finally reopened US embassy like 5 times a day for several weeks to – by luck – catch an interview spot with less than 2 months waiting time while the average wait time was about two years…).

Against all odds and with lots of help from all sides, especially CMU, I was able to get into the US in the midst of the pandemic in June 2021. And just in time, as the Endesfelder Lab was about to move across the Atlantic again (Ulrike had meanwhile accepted a professorship at Bonn University to start on August 1^st, 2021), and this was the last chance to use the microscope before the next moving down time. Luckily, the transfer of strains and protocols to Pittsburgh worked seamlessly and we were able to finish all needed measurements. Again, the timing was crucial: Our Yersinia samples were the very last thing measured at the CMU and literally five minutes later, the group started to disassemble the microscope again to be loaded into the sea freight container only three days after. 

Back in Germany, work continued to get biochemical proof what we saw in the SMLM measurements (as rightfully many other scientists in the field were skeptical that what we saw was true, as the standard attempts so far had not revealed these interactions). Supported by an awesome playlist playing on repeat, my colleague Corentin Brianceau and I aimed to perform the best co-immunoprecipitation ever done on the fickly cytosolic complexes. The weak but consistent results were enough to finalize the manuscript.

It was clear that to truly convince a broad and highly active research community of our new effector shuttle concept was not easy. Luckily, we could convince Francois Meyer, editor at Nature Microbiology, that the question addressed by our study is central to the field, as stated in several major reviews by colleagues. The reviewers trusted our results, but since our approach using SMLM to answer this question is a novel one, they asked for additional proof by other, more established methods. This basically put us back to where we started years ago and to the question how to measure transient interactions of diverse interaction partners within the complex environment of a living cell.

We decided to try all the different approaches we could think of – both in vitro, trying again to purify components, and using another novel in situ approach that had become available within the last years and caught our attention: proximity labeling. Two PhD students, Corentin Brianceau and Katherine Pintor, as well as Jan Vielhauer in his Master’s thesis focused on this project, with expert support by , our technical assistant. They adapted purification protocols, used biolayer interferometry and crosslinking approaches, and established proximity labeling in Yersinia. Alexander Balinovic and Katherine Pintor performed additional SMLM experiments at the University of Bonn in the lab of Ulrike’s next door colleague Ulrich Kubitscheck, who kindly offered optics lab space to her group, as the reconstruction of the Endesfelder lab was – and still is – not finished). Despite various challenges, the additional results were in line with our model and after more than half a year (!!!) of frenzied revision work, we were able to submit a vastly revised version of our manuscript.

Ultimately, after more than six years of work, with several generations of young scientists joining the lab and others (including myself) having moved on, the combined strength of all authors involved and the combination of all these approaches was required to show that effector proteins directly interact with the T3SS cytosolic components in the bacterial cytosol. We could provide both, qualitative and quantitative support for an effector shuttling mechanism from the bacterial cytosol towards the injectisome. Being proud of our endurance and a collaboration that continued with a strong scientific spirit throughout all these years, we think that we provided a truly important conceptual contribution to the bacterial secretion field. Much more generally, we have thoroughly and successfully shown that SMLM and proximity labeling are powerful “in situ biochemistry” methods to investigate complex and dynamic systems in living cells. We believe that techniques like those will yield many more important findings in not only microbial cell biology but generally in cell biology in the future and would like to invite the life science community to try them out for their research questions as well!