Daniel F. Azar (IMBA-GMI Communications & Partnerships): Dear Julius, Ves and Changwei, congratulations to you and your team on this new publication and the exciting findings! In this paper, you find a missing link between the piRNA pathway and the heterochromatin machinery in Drosophila. Were you aware of the existence of such a link prior to making these findings?
Julius Brennecke (IMBA Group Leader): Thanks, Daniel. Sure, we knew that the nuclear piRNA pathway at some point must connect to the basic heterochromatin biology, but we did not know how exactly. A well-established path for specific heterochromatin formation is through DNA-binding factors. However, their adaptive potential for targeting fast-evolving, active transposons is limited. This is where Argonaute-small RNA complexes come into play, precisely due to their high adaptability and versatility. Our lab studies the piRNA pathway in flies, a highly efficient RNA interference pathway that is dedicated to the silencing of transposable elements. Nuclear Argonaute-small RNA complexes act at the level of nascent mRNAs and therefore silence transposons as soon as they start “talking”, i.e., transcribing. We knew that, very upstream, an Argonaute protein with a sequence-complementary small RNA targets nascent transposon transcripts, and that this eventually leads to potent heterochromatin formation and silencing. I find this a fascinating process, as it requires transcription to a certain extent! However, what is between the upstream event (Argonaute binding) and downstream event (heterochromatin formation) is poorly understood.
DA: That means, in a way, a highly adaptable mechanism such as the piRNA pathway functions as a specific signal for the less adaptable, but highly effective chromatin packaging machinery to reach a synergistic effect. And initiating this synergistic effect is a potent transcriptional repressor in the piRNA pathway, Panoramix, that strangely shows features of transcriptional activators…
JB: Exactly! We previously described the SFiNX complex and its role in co-transcriptional silencing downstream of the Piwi Argonaute protein. One component of the SFiNX complex, Panoramix (Panx), is an orphan protein with unknown functions. Panx has a structured C-terminal part, but the N-terminal half of the protein sequence is disordered. Now if we look more closely, we see that this intrinsically disordered region of Panx contains a 200 amino acid (aa) stretch that is highly acidic and rich in prolines. Although Panx is an undisputed transcriptional repressor, these features are reminiscent of transcriptional activators and were called “negative noodles” or “acidic blobs” in that field. In our case, we experimented with this acidic noodle of Panx and identified a 27 aa peptide that confers very strong silencing activity when recruited to a reporter gene. Remarkably, this peptide contained a central, Leucine rich motif (the LxxLL motif), which again was known from the transcription field as a motif that mediates interactions between transcription factors and co-activators.
DA: And identifying this peptide in Panx allowed you to make a breakthrough finding on how the piRNA pathway links to the general heterochromatin machinery… Can you tell me how this happened?
Veselin Andreev (PhD candidate in the Brennecke lab and co-first author): Precisely. We used synthetic peptides containing the LxxLL peptide of Panx, or a mutated version of it, to do a “fishing experiment” by incubating them with nuclear extracts. To our great surprise, this allowed us to identify Sov (“Small ovaries”) as a Panx binding partner! Sov immediately caught our attention because it had a documented role in transposon repression and heterochromatin biology. But Sov is an imposing 400 kilodalton multi-zinc finger protein, certainly one of the last proteins we would have expected to interact with Panx. Looking at Sov, one must assume that it directly binds to DNA with all its zinc fingers (21 in total!). I am not saying that it doesn’t. However, we found a very interesting interaction mechanism between Sov and Panx. The molecular basis can be summarized as follows: the LxxLL motif in the Panx N-terminal acidic noodle interacts with a tri-helical domain in the N-terminus of Sov. But this structural interaction is only one side of the coin! In fact, mass spectrometry experiments allowed us to detect SUMOylation, a post-translational modification (PTM), on that region of Panx. In turn, this alerted us to the presence of SUMO-Interacting Motifs (SIMs) flanking the Sov tri-helical domain. Hence, we see an interesting dual mode of interaction: one weak constitutive structural interaction, and one reversible interaction in the context of a PTM.
JB: Here it is crucial to note that Panx, an essential silencing protein in the piRNA pathway, is essentially inert if it were not SUMOylated. Think of it this way: the ability to SUMOylate at need to strengthen a weak interaction allows to restrict the interaction to the desired place of action. Hence, regulation is the key that allows us to rethink the silencing mechanism. In fact, we could show that Panx SUMOylation only occurs in the presence of Piwi and that it is restricted to chromatin. Many people in the field, including us, have been trying to find a direct interaction between Piwi and Panx or SFiNX for years. It might not exist in the end as the SUMO-dependent regulatory mechanism would not require it. But slowly, the pieces of the puzzle are starting to come together!
DA: Panx being an orphan protein makes its interactions all the more interesting to explore. What can you say about the N-terminal tri-helical domain of Sov? Does this domain fall into any known category in terms of its sequence or fold?
Changwei Yu (postdoctoral fellow in the Brennecke lab and co-first author): We did find an interesting evolutionary origin of this Sov domain in a transcriptional coactivator, Mediator 15 (Med15). This is a well conserved protein among animals, which strongly suggests that during evolution, this domain was “stolen” from Med15 and repurposed in the context of the Sov multi-zinc finger into a transcriptional repressor module. This brings us back to the striking parallels we could find between the transcriptional repressors Panx and Sov and the molecular principles known from transcriptional activators!
DA: It must be exciting to witness how evolution repurposes available units and changes their allegiances! Are we here looking at a transcriptional repression mechanism that has roots in transcriptional activation?
JB: Overall, the molecular principles are copied, but the specificity must have swapped along the way. The mode of interaction itself is agnostic of whether this works as an activator or repressor, as that information is encoded elsewhere. Our new findings, therefore, lead us to another question or speculation: does the massive multi-zinc finger protein Sov utilize its N-terminal domain only for the piRNA pathway? Or is this domain also being used by other silencing systems… It might even be that the Sov-Panx interaction is the more recent invention and that Sov has an older function in specifying heterochromatin downstream of other processes. There are many such examples of ‘molecular repurposing’ in the piRNA field or the broader field of genome defense. And it is one of our future visions to understand this molecular arms race between the gene expression machinery, the transposons, and their repressive systems!