Biological liquid-liquid phase separation gives rise to dense protein-protein or protein-RNA condensates that are distinct from the surrounding bulk cytoplasmic or nuclear phase. These condensates, comprised of many multivalent, weak, and hydrophobic interactions, perform a wide variety of physiological functions and are sensitive to changes in the cellular environment (Shin and Brangwynne, 2017). One notable phase-separated condensate is the P granule, a C. elegans germline-specific mRNA surveillance center. While the liquid nature of P granules was first described by Brangwynne et al. (2009), additional P granule properties and protein dynamics have been examined with a variety of in vivo techniques. The advances in CRISPR/Cas9 gene editing techniques make it possible to endogenously tag PGL-1, a major constituent of P granules, and study protein dynamics in vivo via live fluorescent imaging. PGL-1 tagged with Green Fluorescent Protein (GFP) is widely used for the study of P granules, and forms distinct perinuclear germline foci consistent with previous observations of P granules (A, top row) (Pitt et al., 2000; Strome and Wood, 1982).
Due to an interest in visualizing additional germline proteins, we created both
pgl-1::mKate2 (A, middle row) and
pgl-1::mTagBFP2 (A, bottom row) endogenously tagged strains that would allow us to visualize multiple fluorescent proteins with differing excitation and emission spectra. All three C-terminally fluorescently tagged PGL-1 strains behave similarly to one another under both undissected and dissected control conditions (A, B). However, when we tested conditions that probe P granule properties, we were surprised to find that the mKate2 and mTagBFP2 tagged PGL-1 have strikingly different dynamics from the previously described PGL-1::GFP.
First, 1,6-hexanediol is an aliphatic alcohol that is shown to dissolve liquid-like condensates, but not aggregates or solids (Kroschwald et al., 2017). While the mechanism is not entirely understood, 1,6-hexanediol is thought to disrupt hydrophobic interactions that are important for phase-separated condensate integrity. Consistent with Updike et al. (2011), PGL-1::GFP dissolves in 5% 1,6-hexanediol, as demonstrated by lack of perinuclear foci and increased fluorescent signal in the cytoplasm (C, top row). In contrast, although PGL-1::mKate2 and PGL-1::mTagBFP2 have increased cytoplasmic fluorescent signal, clear perinuclear puncta are still visible, indicating a lack of complete dissolution and perhaps a more aggregate-like consistency (C, middle row, bottom row).
Second, liquid phase separation is influenced by changes in temperature. Higher temperatures introduce greater amounts of entropy into the system, allowing for de-mixing of the condensate into the bulk phase (Alberti et al., 2019). PGL-1::GFP condensates in embryos are observed to dissolve at temperature shifts to 34C for 1 minute (Putnam et al., 2019). In adult germlines, evidence of P granule de-mixing can be seen by 3 hours at 34C, where PGL-1::GFP foci are faint and fluorescent signal is heavily cytoplasmic (D, top row). However, this same temperature shift results in large, bright, abnormal foci and low cytoplasmic signal in both PGL-1::mKate2 and PGL-1::mTagBFP2 (D, middle row, bottom row). The majority of these foci appear detached from the nuclear periphery and large, round, abnormal aggregates are also observed in the syncytial gonad rachis (not shown).
Lastly, P granules are protein-RNA condensates, and rely on RNA for their formation. Consistent with observations by Sheth et al. (2010), PGL-1::GFP foci in the pachytene are disrupted 5 hours post-microinjection of -amanitin, a potent transcriptional inhibitor (E, top row). PGL-1::mKate2 and PGL-1::mTagBFP2 both fail to dissolve, and instead form bright puncta that are abnormal compared to control foci (E, middle row, bottom row).
Both mKate2 and mTagBFP2 are bright, monomeric fluorescent tags derived from TagRFP and are similar in size to GFP (Shcherbo et al., 2009; Subach et al., 2011). Here we show that the in vivo dynamics of PGL-1 foci are drastically altered when tagged with mKate2 or mTagBFP2 and appear to be more solid or aggregate-like than when tagged with GFP. While the mechanism behind this effect is unclear, it is possible that specific tag-to-tag or tag-to-protein interactions are altering the phase dynamics. Additionally, the high expression levels of PGL-1 may make it more prone to aggregation. Ultimately, our data shows fluorescent tag choice is sufficient to perturb phase-separation dynamics, and we recommend ensuring that the dynamics of new fluorescently tagged proteins are consistent with previous literature, in vitro experiments, untagged protein, or additional fluorescent tags.