Genomic tagging of endogenous human ESCRT-I complex preserves ESCRT-mediated membrane-remodeling functions

The endosomal sorting complexes required for transport (ESCRT) machinery drives membrane scission for diverse cellular functions that require budding away from the cytosol, including cell division and transmembrane receptor trafficking and degradation. The ESCRT machinery is also hijacked by retroviruses, such as HIV-1, to release virions from infected cells. The crucial roles of the ESCRTs in cellular physiology and viral disease make it imperative to understand the membrane scission mechanism. Current methodological limitations, namely artifacts caused by overexpression of ESCRT subunits, obstruct our understanding of the spatiotemporal organization of the endogenous human ESCRT machinery. Here, we used CRISPR/Cas9-mediated knock-in to tag the critical ESCRT-I component tumor susceptibility 101 (Tsg101) with GFP at its native locus in two widely used human cell types, HeLa epithelial cells and Jurkat T cells. We validated this approach by assessing the function of these knock-in cell lines in cytokinesis, receptor degradation, and virus budding. Using this probe, we measured the incorporation of endogenous Tsg101 in released HIV-1 particles, supporting the notion that the ESCRT machinery initiates virus abscission by scaffolding early-acting ESCRT-I within the head of the budding virus. We anticipate that these validated cell lines will be a valuable tool for interrogating dynamics of the native human ESCRT machinery.


KI HeLa A ATGGCGGTGTCGGAGAGCCAGCTCAAGA*TGAGGCTGCGACGCGCTCGCCTCCCAG M A V S E S Q L K M R L R R A R L P
* 15 nt deletion removes splice site.

KI HeLa B ATGGCGGTGTCGGAGAGCCAGCTCAAGAAAATGGTGTGCCAAGGTGAGGCTGCGACGCGCTCGCCTCCCAG M A V S E S Q L K K M V C Q G E A A T R S P P
1 nt insertion causes frameshift.

nt insertion
introduces premature stop codon.

nt insertion
introduces premature stop codon.

Figure S2. GFP-Tsg101 knock-in cell lines have knockout indels on their untagged Tsg101 alleles.
To sequence the non-knock-in alleles of Tsg101, PCR was performed on genomic DNA of the knock-in cell lines with primers that anneal to the genomic DNA on either side of the Tsg101 target site. The primers could also anneal to a knock-in allele, but that would give a much larger PCR amplicon, which was excluded by limiting the extension time and by gel-purifying the band at the predicted size for a non-knock-in allele. Sanger sequencing of the PCR products showed indels that knock-out expression of Tsg101. The indels, and their effects on the translated protein sequence, are indicated in red. This confirms that the knock-in cell lines have no functional alleles for untagged Tsg101.    Figure S3. GFP-Tsg101 knock-in alleles show precise insertion. (A) 5' and (B) 3' homology junctions of the KI alleles were amplified by PCR from genomic DNA, analyzed by Sanger sequencing, and compared to the human genome reference sequence, donor plasmid sequence, and sequences from the parental (WT) cell lines. Non-coding regions (5'UTR and intron) are shown in black/white text, Tsg101 exon 1 in blue, PuroR in red, and GFP in green. The results show precise integration of the PuroR-P2A-GFP insert, with no abnormalities except for several single-nucleotide polymorphisms in non-coding regions, some of which were also found in the parental lines. The discrepancies in the Tsg101 coding sequence in the alignment are the silent mutations intentionally introduced to the donor plasmid to prevent CRISPR from cutting the knock-in allele. Figure S4. GFP-Tsg101 knock-in cell lines express GFP-Tsg101 with no free GFP. Western blotting for GFP in cell lysates showed that all GFP in the KI cell lines was the GFP-Tsg101 fusion protein, with no free GFP detected. The additional bands, which were detected in the wild-type (WT) parental lines as well as in the GFP-Tsg101 KI lines, represent nonspecific binding of the antibody. Figure S5. Viral release assays performed via transduction confirm that HIV-1 release and its dependence on Tsg101 are not disrupted in GFP-Tsg101 KI cell lines. As an alternative approach for assessing HIV-1 particle release, HeLa and Jurkat cells were transduced as described in the methods. (A) Virus release assays, as described in Fig. 5  The distribution of single molecule GFP photon counts was fit to an exponential curve to calculate upper and lower bounds (± 1 s.d.) around the mean: single GFP molecule = 231-389 photons. These upper and lower bounds for the numbers of photon counts per GFP molecule were used to calculate the lower and upper bounds respectively for the numbers of GFP molecules per HIV-1 particle in Fig. 6D.