Original article: Trentini, D.B.,Suskiewicz, M.J.,Heuck, A.,Kurzbauer, R.,Descz, L.,Mechtler, K.,Clausen, T. "Arginine phosphorylation marks proteins for degradation by a Clp protease." Nature (2016) doi:10.1038/nature20122
Bacillus subtilis has a very special smell. Some say it’s like dirty socks, I think it’s more like moist soil, but anyway, when you study an organism, you learn to “accept it as it is." And, additionally, it’s not like we grow it every single day. Our lab – although relatively small, with a dozen-odd members including Tim – works on a few different organisms, and when we need proteins to study in vitro, we usually produce them in E. coli or insect cells.
On those days, however, the dominant smell – filling the main lab and the cold room, lingering in the corridors – was that of B. subtilis. Débora, a great colleague and the mass spec/biochemistry expert on the project, was as busy as a bee and the expression on her face said she was not to be distracted. In this series of experiments her task consisted each time of six parts performed in parallel – three repeats of the actual experiment and three controls – so there was much to do on one day and no room for a mistake.
The idea was to analyse, using quantitative mass spectrometry, whether substrates co-purified with ClpP are phosphorylated on arginine residues. ClpP – a barrel-shaped homoheptameric structure with active sites facing inwards – is the major protease of B. subtilis, the bacterial counterpart of the 20S proteasome responsible for the bulk of intracellular protein degradation. We hypothesised that the role of a “bacterial ubiquitin”, a degradation tag that targets substrates into ClpP, might be fulfilled by arginine phosphorylation. This modification was discovered several years ago by the joint efforts of our group and Karl's mass spec team on campus and so far confirmed only in gram-positive bacteria. If our prediction was correct, we should be able to see phosphoarginine on proteins trapped inside ClpP. And so it started.
Based on the literature, Débora designed a mutant ClpP variant that would bind substrates but not hydrolyse them, and she expressed its tagged version in B. subtilis. First she did it in a clpP-deficient strain, to prevent the formation of mixed heptamers that would contain both active and inactive subunits. But – consistent, in fact, with the hypothesis we wanted to test – knocking out clpP resulted in a global accumulation of phosphoarginine-carrying proteins in the cell. It was clear that, to get an unbiased result, we had to use cells that are WT for clpP, but at the same time ensure that the endogenous subunits and the plasmid-expressed subunits don’t mix. Here was a puzzle.
The solution was provided by what I think was one of the breakthroughs in the project – Tim’s idea to perform salt bridge swapping at the heptameric interface of ClpP that prevented the interaction between the two ClpP versions. The experimental set-up was ready. Débora spent many days doing the pull-downs and saw that, sure enough, phosphoarginine-tagged proteins are reproducibly co-purified with ClpP.
The rest of the project was the extension of this initial finding. We confirmed soon after that the mechanism required also ClpC, an unfoldase that – much like the regulatory subunit of the proteasome in the eukaryotic ubiquitin-dependent system – sits on top of the protease barrel and feeds in appropriately tagged substrates. These two proteins, ClpC and ClpP, plus, of course, arginine phosphorylation on the substrate, was all that was needed for the mechanism to work. Commercially supplied β-casein, when phosphorylated on arginine residues (using the protein arginine kinase McsB) and then mixed with pure ClpC, ClpP, and Mg-ATP, was degraded within minutes. This part of the project – the many degradation assays testing which components are needed and eventually yielding the minimal system just described – had a special beauty to it, one that only in vitro enzymology can have, I think. Some say biochemistry is in retreat, but we believe that Efraim Racker’s “Don't waste clean thinking on dirty enzymes” still holds.
The final confirmation of the mechanism came from the co-crystallisation of the N-terminal domain of ClpC with the amino-acid phosphoarginine – and here forgive me a brief digression. I remember very well when, a few years earlier, I came to Vienna for PhD interviews. Each group leader presented their project and so did Tim, his talk starting with a picture of the phosphoarginine side chain and a question to the audience: “Do you know what this molecule is?” At that time, it wasn’t very clear what phosphoarginine does in the cell, but I thought – as I would now – that it was fascinating enough to study a novel post-translational modification that had hundreds of in vivo substrates (as I learnt only on that day, as it was still unpublished). Now, a few years after, I saw the same molecule but not as a drawing – rather, as an electron density, filling a specific cleft in ClpC. The cleft had a positively charged and a negatively charged part, exactly matching the unique charge distribution of the phosphoarginine side chain. Based on the structure, we could design a specific mutant of ClpC that was deficient in the degradation of phosphoarginine-carrying proteins but not in its other functions.
The lab was again filled with the familiar smell of B. subtilis – we had to test whether the system is important in vivo. The clpC-deficient B. subtilis strain showed a clear loss of thermotolerance, likely because the cell couldn’t degrade misfolded proteins that accumulate upon heat. The phenotype was complemented by expressing WT ClpC, but not the specific mutant that was incapable of recruiting phosphoarginine-carrying proteins. The experiment was consistent with reports from other gram-positive bacteria – including Staphylococcus aureus and Bacillus anthracis – which saw the unfoldase ClpC and the protein arginine kinase McsB to be important for stress resistance and virulence.
For us, the study will remain first and foremost a most beautiful adventure in basic biology. And an adventure in team working, for all the individual contributions – and the countless meetings and discussions we have had, always coming up with more ideas than could be tested – turned out to be important to put the final model together.