Expert Insight: Mapping the human proteome using CRISPR-mediated fluorescence tagging, microscopy, mass spectrometry, and machine learning

---

Dr. Manuel Leonetti, Intracellular Architecture Group Leader at the Chan Zuckerberg Biohub

Proteins are the product of gene expression and the molecular building blocks of cells. But while the genome sequence defines the set of all proteins that make up our cells, a systematic characterization of how the proteome is organized within the cell remains an important goal of modern cell biology. A comprehensive map of the human proteome’s organization will serve as a reference to understand gene function in health and disease.

In this SelectScience^® webinar, now available to watch on demand, Dr. Manuel Leonetti, Intracellular Architecture Group Leader at the Chan Zuckerberg Biohub, describes how his team combined CRISPR engineering, flow cytometry-based cell sorting, confocal live-cell imaging, mass spectrometry, and machine learning to systematically map the subcellular localization and interactions of 1,310 human proteins. Their approach provides a data-driven description of the molecular and spatial networks that organize the proteome.

Read on for highlights of the Q&A session, and register now to watch on demand.

How can we make sure that when we fuse a given protein with green fluorescent protein (GFP) we're not perturbing its function?

ML: The main message is GFP tagging is a tool in an approach that is absolutely safe, and it has been tested so much that we know we can use it in very physiological contexts.

But specifically, the way we think about this question is twofold. Firstly, there are entire proteomes, especially yeast, that have already been tagged with GFP. So, we have a lot of data that show that yeast strains, for example, are completely functional.

Secondly, we can look specifically, and we did that in our cell lines. What happens when we tag an essential gene? These are proteins and genes that are so important for the life of a cell that even if we messed up their function a little bit, we should see a fitness defect. We looked at that and don't see any fitness defects in our approach.

The most important part is that we're very careful at choosing the insertion sites when we want to tag a specific protein with GFP. We do three different things. We look at the literature first for every protein to find studies that show that insertion of GFP at one specific terminus is safe to preserve protein function. Number two, we check if there is structural information available for a specific protein, especially if that protein is known to be part of a complex; we check the protein data bank and look at whether the insertion site is exposed. If it's buried, we don't want to insert something in an important protein-protein interface.

Finally, we check for the presence of any important regulatory sequences in the sequence of a specific protein and make sure that we don't insert anywhere near, for example, a localization sequence. So, there's a lot of work that we do up front to make sure that we choose the insertion site in a way that, as much as possible, it is not going to perturb a protein’s function.

What are the therapeutic implications of mapping human protein using CRISPR technology? Are there ethical problems with its use of human DNA? And does using this technique require regulation to ensure proper use to enhance therapeutic outcomes?

ML: There are a few different layers to that question. Firstly, what are the therapeutic implications of our work? At the core, our work is interested in mapping what is going on in the cell to be able to interpret disease. So, the therapeutic implication is creating knowledge that we hope is going to allow us to develop therapies to understand the mechanisms of disease faster and, therefore, develop therapies faster.

There are a lot of mutations that we find in patients, for which we have no idea how they link to specific disease phenotypes. In some of these cases, we think that it's because we don't understand the function of the different proteins that are implicated in these mutations and how these proteins communicate with other parts of the cell. So, having a comprehensive map should really help us disentangle what's going on here.

The second implication for therapeutic applications is that all the work that we're doing, especially with microscopy, are great setups for drug screening. We're hoping that the cell lines that we're building could be used for high-throughput drug screening and be able to find molecules that have specific bioactivities.

The second part of the question is about how CRISPR is going to be used in gene editing, particularly in patients. We don’t work on this; we work specifically with in vitro systems for research questions. In terms of the regulatory questions, etc., that are related to all of these, there are already drugs in the clinic that use CRISPR technology, and these are completely safe.

A lot of the hope of CRISPR-based therapy is taking cells directly from a patient, repairing what is wrong with these cells ex vivo, and then putting these cells back into the same patient. This is being done, for example, in immuno-oncology. There are already clinical trials on the way where if you have a specific cancer, I could take some of your immune cells or your T cells, CRISPR in a specific CAR-T receptor so that your T-cells would attack the specific kind of cancer that you have, and then put these cells back into your body.

There's this idea of tailoring things to a specific patient, which makes sense from a regulatory perspective. It is safe. There is no involvement of using cells that are derived from somebody else, etc., so there's a very complex set of therapies that are being developed with CRISPR. I think the regulatory questions that are being thought about right now are being solved. A lot of these drugs are already in the clinic. They've shown potency, and I think this is going to be very exciting for patients.

Want to learn more about mapping protein-protein interactions? Watch the webinar on demand>>

SelectScience runs 10+ webinars a month, discover more of our upcoming webinars>>