Experimental considerations for study of C. elegans lysosomal proteins

Abstract Lysosomes are an important organelle required for the degradation of a range of cellular components. Lysosome function is critical for development and homeostasis as dysfunction can lead to inherited genetic disorders, cancer, and neurodegenerative and metabolic diseases. The acidic and protease-rich environment of lysosomes poses experimental challenges. Many fluorescent proteins are quenched or degraded, while specific red fluorescent proteins can be cleaved from translational fusion partners and accumulate. While studying MLT-11, a Caenorhabditis elegans molting factor that localizes to lysosomes and the cuticle, we sought to optimize several experimental parameters. We found that, in contrast to mNeonGreen fusions, mScarlet fusions to MLT-11 missed cuticular and rectal epithelial localization. Rapid sample lysis and denaturation were critical for preventing MLT-11 fragmentation while preparing lysates for western blots. Using a model lysosomal substrate (NUC-1), we found that rigid polyproline linkers and truncated mCherry constructs do not prevent cleavage of mCherry from NUC-1. We provide evidence that extended localization in lysosomal environments prevents the detection of FLAG epitopes in western blots. Finally, we optimize an acid-tolerant green fluorescent protein (Gamillus) for use in C. elegans. These experiments provide important experimental considerations and new reagents for the study of C. elegans lysosomal proteins.


Introduction
Lysosomes are membrane-enclosed cytoplasmic organelles required for the degradation of diverse biological macromolecules (Ballabio and Bonifacino 2020). Consistent with this function, they are among the most acidic compartment in the cell with a pH ranging from 4.5 to 5.5, and are packed with proteases, nucleases, acid lipases, and carbohydrate processing enzymes (Bonam et al. 2019). Lysosome dysfunction can lead to inherited lysosomal storage disorders, as well as neurodegenerative and metabolic disease, and cancer (Ballabio and Bonifacino 2020). Lysosome activity declines with age and is required for lifespan extension (Hansen et al. 2008;Sun et al. 2020).
Using fluorescent protein (FP), fusions to study lysosomal lumen proteins present challenges. Many green and red FPs derived from avGFP and eqFP578, respectively, are sensitive to degradative lysosomal proteases (Shinoda, Shannon, et al. 2018). The sensitivity of many other FPs to lysosomal proteases remains to be determined (Shinoda, Shannon, et al. 2018). Due to their low pKa (3.1-5.3) and resistance to lysosomal proteases, red FPs derived from DsRed or eqFP611 (i.e. mCherry, mScarlet, and mRuby) are typically the FP of choice for imaging lysosomal lumen proteins (Shinoda, Shannon, et al. 2018). An additional consideration in interpreting lysosomal localization is that lysosomal proteases can cleave flexible linkers or the N-terminus of FPs, separating the FP from the protein of interest (Ko et al. 2003;Kollmann et al. 2005;Huang et al. 2014;Miao et al. 2020). While this cleavage can be used to monitor lysosomal activity (Miao et al. 2020), it can hamper interpretation of lysosomal localization of fusion proteins. Another issue is that many FPs lose fluorescence in the acidic lysosome through fluorophore quenching due to their neutral pKa (Shinoda, Shannon, et al. 2018). Acid-tolerant green FPs have been recently developed but have not yet been widely adopted (Roberts et al. 2016;Shinoda, Shannon, et al. 2018).
During the course of studying MLT-11, a putative Caenorhabditis elegans protease inhibitor, we used CRISPR/Cas9 to introduce an mScarlet::3xMyc tag into the endogenous mlt-11 locus to produce a C-terminal translational fusion that should label all isoforms (Ragle et al. 2022). This strain displayed robust MLT-11:: mScarlet::3xMyc localization in punctae and tubules reminiscent of lysosomes. However, we were unable to verify the fusion was full length by anti-Myc western blotting. We also generated an equivalent MLT-11::mNeonGreen::3xFLAG fusion without a linker (Ragle et al. 2022), which displayed similar punctate/tubular localization, but also transient cuticular localization. This discrepancy between these strains motivated us to explore whether we could minimize cleavage of the FP fusion and explore acid-tolerant green FPs for lysosomal translational fusions.

Strains and culture
Caenorhabditis elegans were cultured as originally described (Brenner 1974), except worms were grown on MYOB media instead of NGM. MYOB agar was made as previously described (Church et al. 1995). We obtained wild-type N2 animals from the Caenorhabditis Genetics Center (CGC). Table 1 contains a list of all the strains used and their source.

Transgenesis and genome editing
All plasmids used are listed in Supplementary Table 1 Table 1 ;. The mCherry cassette in pJW2139 was replaced with mScarlet-I through Gibson cloning to create pJW2145. The hsp-16.41::nuc-1::linker::Gamillus::linker::red FP::tbb-2 3′UTR cassettes from pJW2139 and pJW2145 were PCR amplified and SapTrap cloned (Schwartz and Jorgensen 2016) into pNM4216 to generate pJW2460 and pJW2461, respectively. pNM4216 is an insertion vector for rapid recombination-mediated cassette exchange (rRMCE). rRMCE is a derivative of RMCE that generates usable knock-ins more quickly by removing the need to excise a selectable marker (see https://sites.wustl.edu/ nonetlab/rapid-rmce-beta-testing/, last updated 2022 June 15 for more details and for protocols). pJW2460 and pJW2461 were integrated into NM5548 using rRMCE to generate strains JDW556 and JDW557, respectively.

Microscopy
Animals were picked into a 5 µl drop of M9 + 0.05% levamisole solution on a 2% agarose pad on a microscope slide, then a coverslip  Schindelin et al. 2012). For direct comparisons within a figure, we set the exposure conditions to avoid pixel saturation of the brightest sample and kept equivalent exposure for imaging of the other samples. For co-localization analysis, animals of the indicated genotype were synchronized by alkaline bleaching (dx.doi.org/ 10.17504/protocols.io.j8nlkkyxdl5r/v1), released on MYOB plates, and incubated at 20°C for 48 h. Plates were heat-shocked at 34°C for 30 min and then incubated at 20°C for an additional 24 h. Animals were picked and imaged with a 63 × objective, as described above except no agarose pad was used for image acquisition. Consistent exposure times for green and red FP imaging were used for each strain. Background was removed using a rolling ball method in Fiji (radius = 50 pixels). Subsequent analyses were performed using Imaris software (Oxford Instruments). A mask was created using surface detail = 10 microns, voxel intensity = 10. The Coloc tool was then used with a threshold set to 0.25 and PSF width set to 0.25. A Mander's test was performed on the co-localization data (Manders et al. 1992(Manders et al. , 1993.
For all other western blots, 40 animals were picked into 40 µl of M9 + 0.05% gelatin and Laemmli sample buffer was added to 1 × and then immediately incubated for 5 min at 95°C. Lysates were then stored at −80°C until they were resolved by SDS-PAGE. For the western blots in Fig. 3, animals were synchronized by bleaching and harvested at the indicated times. Lysates were resolved using precast 4-20% MiniProtean TGX Stain-Free Gels (Bio-Rad) with a Spectra Multicolor Broad Range Protein Ladder (Thermo; # 26623) protein standard. Proteins were transferred to a polyvinylidene difluoride membrane by semi-dry transfer with a TransBlot Turbo (Bio-Rad). Blots and washes were performed as previously described & nbsp; (Ragle et al. 2020). Anti-FLAG blots used horseradish peroxidase-conjugated anti-FLAG M2 (Sigma-Aldrich, A8592-5 × 1MG, Lot #SLCB9703) at a 1:2,000 dilution. Mouse anti-alpha tubulin 12G10 (Developmental Studies Hybridoma Bank; "-c" concentrated supernatant) was used at 1:4,000. Rabbit anti-mCherry (AbCam ab167453) was used at 1:1,000. The secondary antibodies were Digital anti-mouse (Kindle Biosciences LLC, R1005) diluted 1:20,000 or Digital antirabbit (Kindle Biosciences LLC, R1006) diluted 1:1,000. Blots were incubated for 5 min with 1 ml of Supersignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific, 34095) and the final blot was imaged using the "chemi high-resolution" setting on a Bio-Rad ChemiDoc MP System.

Sample processing affects MLT-11::mNG stability in lysate generation
As fluorescent tags can be cleaved off fusion proteins in lysosomes (Miao et al. 2020), western blotting to confirm that a fusion protein is full length is essential to have high confidence in lysosomal localization. For our mlt-11::mScarlet::3xMyc strain, we were never able to detect bands of the predicted size by western blotting with anti-Myc or anti-mScarlet antibodies (unpublished data). When attempting western blots on mlt-11::mNeonGreen::3xFLAG lysates, we saw variable laddering (e.g. see Fig. 2b, lane C). As lysosomal proteases can degrade proteins, we sought to optimize our sample preparation conditions to minimize degradation. We harvested samples at peak MLT-11 protein expression (42 h postrelease, stage L4.3 (Mok et al. 2015;Ragle et al. 2022), and tested a range of variables: (1) denaturation at 70°C for 10 min vs 95°C for 5 min; (2) denaturing samples immediately after collection vs rapid freezing and denaturation of all samples together later; (3) rapid freezing using dry ice vs liquid nitrogen; and (4) whether it was better to denature before storage at −80°C vs denature immediately before resolving samples by SDS-PAGE (Fig. 2a). The best approach was to harvest animals, add Laemmli sample buffer and immediately denature before storage at −80°C (Fig. 2b,  lanes D and H). Denaturation at 95°C for 5 min produced less laddering than heating to 70°C for 10 min (Fig. 2b, compare D to H). The other approaches with various combinations of rapid freezing and denaturation all produced more degradation products above 50 kDa (Fig. 2b). In all conditions, there is a strong band at 50 kDa (Fig. 2b), consistent with a C-terminal MLT-11 fragment we previously observed (Ragle et al. 2022). These experiments demonstrate that sample preparation has a significant effect on MLT-11 stability during the preparation of lysates for immunoblotting.

Tag cleavage is not reduced by proline linkers or truncated mCherry
As red FPs are stable in lysosomes and linkers can be cleaved by lysosomal proteases (Shinoda, Shannon, et al. 2018), red FP accumulation might not reflect true localization of a fusion protein.

The NUC-1 FLAG epitope is not recognized in immunoblotting after extended time in the lysosome
In our FLAG immunoblots, the full-length NUC-1::P5:: crmCherry::3xFLAG product declined in intensity in late-L4 and adult animals and we did not observe a band at the expected cleavage product position (Fig. 3a). In contrast, in the anti-mCherry immunoblots the cleavage product increased in intensity in late-L4 and adult animals. mlt-11 mRNA levels oscillate and the promoter shuts off in mid-L4, so we are monitoring NUC-1::mCherry and NUC-1::P5::crmCherry::3xFLAG produced by the last pulse of gene expression driven by the mlt-11 promoter (Frand et al. 2005;Hendriks et al. 2014;Meeuse et al. 2020). These data suggest that the FLAG epitope is not recognized in the cleaved mCherry fragment, an important consideration in interpreting anti-FLAG immunoblots.

The green FP Gamillus is not quenched in C. elegans lysosomes
Another limitation of FP usage in lysosomes is that many green FPs are quenched and degraded (Shinoda, Shannon, et al. 2018). The quenching could produce a false negative for lysosomal expression of a fusion protein. As co-localization studies frequently rely on red and green FPs, we sought alternate green FPs for lysosomal imaging. Two candidates from the literature were pH-tdGFP and Gamillus. pH-tdGFP is an engineered tandem dimer, which is acid-tolerant and stable in vitro over a pH range from 3.75 to 8.50 (Roberts et al. 2016). However, we did not pursue this green FP as the tandem dimer would make it a large insertion for knock-ins, which could decrease editing efficiency. Gamillus is an acid-tolerant monomeric green FP developed through directed evolution of a novel green FP from the flower hat jellyfish, Olindias formosa (Shinoda, Ma, et al. 2018). It has a pKa of 3.4 and is reported to have a useful combination of brightness, photostability, and maturation speed (Shinoda, Ma, et al. 2018). Gamillus is photoswitchable; at its peak excitation wavelength of 504 nm, it is switched to an off state which could be reversed by irradiation with 352-388 nm light (Shinoda, Ma, et al. 2018). Excitation in the 440-480 nm range produced negligible photochromism, potentially due to a higher on-switching rate (Shinoda, Ma, et al. 2018).
To test whether Gamillus fluoresces in C. elegans lysosomes, we created a heat shock-inducible nuc-1::mCherry::Gamillus transgene. Gamillus and mCherry colocalized in lysosomes 24 h post-heat shock including in tubular, acidified lysosomes (Fig. 4, a and b). This result is in contrast to NUC-1::sfGFP::mCherry, where the sfGFP is quenched over time by the acidic lysosomal environment and there is poor co-localization 24 h postheat shock (Fig. 4, a Miao et al. 2020). As the pKa of mScarlet is higher than that of mCherry (pKa 5.3 vs 3.1), we used this approach to test whether mScarlet is quenched by the lysosomal environment. We constructed a heat shock-inducible nuc-1::mScarlet::Gamillus and demonstrated that mScarlet and Gamillus also colocalized, suggesting that mScarlet is not quenched or degraded in the lysosome (Fig. 4, a and  b). Gamillus had a significantly higher co-localization with mCherry or mScarlet by Mander's coefficient in comparison to sfGFP (Fig. 4b).
There was no significant difference in Gamillus co-localization with mCherry and mScarlet (Fig. 4b). These data indicate that Gamillus is a suitable green FP tag for lysosome lumenal proteins and that mScarlet is not quenched in lysosomes.
We next tested whether Gamillus affects the function of proteins to which it is fused, using proteins sensitive to tag dimerization. We used CRISPR to knock Gamillus coding sequence into a histone H3B (his-72) and lamin (lmn-1). We also tagged a germline helicase that localizes to P granules, which are found in ribonucleoprotein condensates. We observed the expected chromatin (his-72), nuclear envelope (lmn-1), and perinuclear (glh-1) localization for each fusion (Fig. 5, a-e). Notably, Gamillus::GLH-1 knock-ins were dimmer than GFP knock-ins (Fig. 5e), consistent with the need to image Gamillus at a wavelength that produces 50% excitation to avoid photoconversion ( Fig. 5e; Shinoda, Ma, et al. 2018). These data suggest that Gamillus does not cause mislocalization and validates the FP for tagging proteins by CRISPR-mediated genome editing. Together, these results validate Gamillus as a green FP option for studying lysosome lumenal proteins.  Images are representative of 20 animals from 2 independent replicates. Scale bars = 20 µm. Gamillus::LMN-1 and DIC image of adult head (c) and embryos (d). Images are representative of 20 animals from 2 independent replicates. Scale bars = 20 µm. GFP::GLH-1 and Gamillus::GLH-1 germline images along with DIC overlays. A Gamillus::GLH-1 image where a 5 × longer exposure was performed is provided. Different animals were imaged for the 1 × and 5 × Gamillus exposures but all images are representative of 20 animals from 2 independent experiments. Scale bars = 20 µm. different issues for green and red FP fusions. Green FP degradation and/or quenching could create false negatives for lysosomal localization. Conversely, the stability of red FPs in the lysosome could allow a cleaved red FP tag to accumulate in the absence of the fusion protein, creating a false positive for lysosomal localization of a factor of interest. Additionally, the bright lysosomal signal can produce high background, obscuring dimmer localization of a translational fusion of a protein of interest in other tissues or cellular compartments. Determining the extent of FP cleavage by western blotting is a critical control to interpret any lysosomal localization of fusion proteins. Sample processing made a major difference in MLT-11::mNG::3xFLAG degradation during western blotting. While the levels of the full-length protein were not obviously affected, the laddering can impair identification of isoforms (Fig. 2). There was a ∼50 kDa band that must be produced by C-terminal cleavage of MLT-11 (Fig. 2) for which we have observed a peak in expression early in L4 (Ragle et al. 2022). We are currently pursuing how this isoform is produced and whether it plays any distinct roles in molting. Our data also suggest that FLAG epitopes become unrecognizable by anti-FLAG antibodies after extended time in lysosomal environments (Fig. 3). We observed a similar phenomenon with MLT-11::mNeonGreen::3xFLAG where in late-L4 larvae and early adulthood we observed lysosomal localization but no signal by anti-FLAG immunoblotting (Ragle et al. 2022). These results are likely due to degradation of the epitope by lysosomal proteases, though we cannot rule out posttranslational modification of the FLAG tag in the lysosome that prevents antibody binding. Using antibodies against FPs may be preferable to use in immunoblotting as a way to test whether the fusion protein is full length. While the rigid proline linker and mCherry N-terminal truncation did not reduce mCherry cleavage from NUC-1, it is possible that they may work on other proteins.

Discussion
We also validated Gamillus as a green FP for labeling the lysosomal lumen and for fusion to lysosomal proteins. When fused to NUC-1, it displayed similar co-localization and acid-tolerance as mCherry (Fig. 4). We also confirmed that despite its higher pKa than mCherry, mScarlet is acid-tolerant making it suitable for lysosomal experiments (Fig. 4). Gamillus also exhibits photoswitching behavior at its peak excitation wavelength. If nonpeak excitation wavelength (440-480 nm) is used the switch to the off-state is minimized at the cost of brightness (Shinoda, Ma, et al. 2018).
We note that our analyses focused on the hypodermis. Hypodermal lysosomes undergo changes in activity, becoming highly active during the molt when they help recycle cuticular components. Intestinal cells are another major source of lysosomal activity in the animal. Gut granules are specialized lysosome-related organelles that play roles in lipid transfer, metabolism, detoxification, signaling, and zinc storage (McGhee 2007;de Voer et al. 2008). Testing the performance of Gamillus, P5 linkers, and western blot sample processing in this tissue is an important area of future exploration.

Data availability
Strains and plasmids are available upon request. To facilitate generation of repair templates and subcloning, all plasmid sequences are provided in Supplementary File 1, knock-in sequences are provided in Supplementary File 2. Sequences of oligonucleotides used for cloning are available upon request.
Supplemental material is available at G3 online.