Deconstructing NgAgo

Review of a Journal Article Published in Nature Biotechnology

By Kate Qin Zhao, PhD

Drafted on 10/20/2016

Authors’ words

In the field of Scientific Research, the key to success is trust.

As an individual, each one of us will try our best to generate quality data in support of a working hypothesis, contributing to the cumulative database of knowledge that may be mined deeper by others later.

We understand that some experiments may not turn out the way we’ve intended; we are not afraid to test something new or something unknown that may lead us to failure. Our own data will be replicated and verified by our colleagues; in return, we will do the same for them. If our initial conclusions are proven to be incorrect, we accept it, acknowledge it, identify the causes as needed, learn our lessons, and move on. We trust others will do the same.

Once the majority of us are practicing these principles, we will have a healthy self-correcting ecosystem. The bond between us will grow stronger, last longer, and together as a community, we will advance our understandings of the world; we will develop technologies to help make it a better place.

The strength of our bond could be challenged when our trusts are tested. Occasionally, desire for monetary reward and personal glorification may lead some to attempt misrepresentation and/or falsification of data. We will grow out of these challenges by working together, by continuing to practice the principles. Our self-correcting ecosystem will evolve, and our bond will grow even more after we, as a community, triumph over these challenges.

Journal Article Review- The Case of NgAgo

This blogpost is a review of a journal article published on NgAgo for gene editing (1).

From this point on, the journal article of interest will be referred to as the NBT paper.

This review contains the following sections

1. What is “One Guide Faithful”?

2. How does NgAgo/gDNA compare to CRISPR/Cas9?

3. Why could others not replicate NgAgo/gDNA for gene editing?

4. Conclusions and recommendations

OK, let’s get started.

Section 1. What is “One Guide Faithful”?

In fiction, there is “One Ring to bring them all and in the darkness bind them.”

In reality, there is “One Guide Faithful” that has kept many scientists bound to NgAgo in the summer of 2016.

According to the NBT paper, “One Guide Faithful” is the characteristic of an efficient gene editing system composed of one protein (NgAgo) plus one guide (5’p ssDNA), ie, NgAgo/gDNA.

NgAgo + 5’p ssDNA = NgAgo/gDNA

(Protein) (Guide)

NgAgo/gDNA + target DNA → Gene Editing

The schematic of “One Guide Faithful” is depicted in Fig.2d of the NBT paper.

Section 2. How does NgAgo/gDNA compare to CRISPR/Cas9, the gold standard for gene editing?

According to the NBT paper, NgAgo is a winner!

Side-by-side comparison of the two systems, based on data presented in the NBT paper as well as the characteristics of each component in the gene editing systems. The x-axis is scaled in relative terms, with NgAgo/gDNA at 100 for each feature compared. The following factors have been considered for this comparison: For efficiency — Data in the NBT paper; For fidelity — Data in the NBT paper and the length of guides targeting specific sequences; For target sequence — Requirement for conserved motifs; For availability — Extra steps in cloning guide sequences; For stability — Features of DNA vs RNA

NgAgo/gDNA has very clear advantages over CRISPR/Cas9 judging by the data presented in the NBT paper and the characteristics of each component in the gene editing systems:

• Higher fidelity (less off-target effect)
• Fewer sequence constraints (no need for PAM)
• Guides are of higher stability (DNA as compared to RNA in CRISPR/Cas9 system)
• Guides are readily available (5’p ssDNA oligo could be purchased at low cost)
• Comparable gene editing efficiency

NgAgo/gDNA could be a potential breakthrough technology for gene editing. However, what’s on paper has not been translated quite well yet in practice, as many are struggling to replicate it (2).

Section 3. Why could others not replicate NgAgo/gDNA for gene editing?

We do not know the exact causal factors behind this “failed to replicate” issue.

The only thing we can do is to take a closer look at the NBT paper, Figure by Figure.

Before we dive into the details of the NBT paper, let’s review the basic requirements for an efficient gene editing system:

1. In the presence of a “guide”

a. It targets specific sequences

b. It acts as a nuclease

c. It makes double strand break (DSB) in DNA

• Example: linearize a plasmid in vitro

2. It is able to access the genome

a. Nuclear localization of the “gene editor”

• Co-localize with genomic DNA

b. in vivo, it generates target specific, genome level indels

Now, let’s check the supporting evidence for NgAgo in the NBT paper:

The authors are quite thorough. They tested 3 different NgAgo fusions:

• GST-NgAgo, purified from E coli
• FLAG-NgAgo-HA, purified from HEK293T cells
• NLS-NgAgo, expressed in mammalian cells

Note: The NLS-NgAgo vector is available from Addgene.

As described by the NBT paper, they’ve checked and validated that NgAgo could be guided to target DNA sequences, act as a nuclease, introduce DSB on target DNA, and generate genome level indels with efficiency similar to CRISPR/Cas9.

Everything fits well for a gene editing system, nothing abnormal.

But wait, let’s simplify the experimental design and look at the data from a different angle.

Experimental design and logic of the workflow in the NBT paper

Hey, something looks interesting here:

There is a jump from needing Two guides to only requiring One guide for NgAgo to make DSB in DNA!

• E. coli expressed and purified GST-NgAgo needs 2 guides to linearize plasmid
• Mammalian expressed and purified FLAG-NgAgo-HA only needs 1 guide to linearize plasmid

Let’s zoom in, and look at this jump again.

Curious jump from 2 guides to 1 guide for NgAgo

Why has this jump happened?

Is it because NgAgo acts differently as a GST-NgAgo fusion than as a FLAG-NgAgo-HA fusion?

Could this be due to different codon optimization of the coding region of NgAgo for each system? (Though there is no mentioning of codon optimization in the NBT paper……..)

Is it because one is purified from E. coli while the other is from mammalian cells?

Do these two NgAgo fusion proteins behave differently?

• How do these fusion proteins look like on an SDS PAGE gel?
• Are there any Western Blots for these proteins in the NBT paper?
• Is one of them truncated or modified?
• Are there any Mass spec data for these proteins to verify their identities?

Since these purified proteins are crucial for generating Fig 1, Fig 2 and Fig 3 of the NBT paper, let’s check the data on the purified proteins carefully.

But, where are the data for purified NgAgo fusion proteins?

These data could not be found in the NBT paper.

OK, now confusions arise due to the lack of data.

We’ve already known that many scientists could not replicate Fig 4 and Fig 5 of the NBT paper, the two key figures for supporting NgAgo/gDNA’s ability in editing a mammalian genome in vivo (see reference 2).

How about Fig 1, Fig 2, Fig 3a, Fig 3b, and also, Supplementary Fig 1, Supplementary Fig 4?

• Collectively, these figures have proven that NgAgo/gDNA meets the requirements of an efficient gene editing system, and, these figures have laid the foundation for Fig 4 and Fig 5.
• The concept of “One Guide Faithful” has first been proposed in Fig 2, then, supported by Fig 3 and Supplementary Fig 1 in the NBT paper.

Can these figures be replicated by others?

Confused?

• Let’s stop for a moment and review the key findings of the NBT paper again.
• Also, let’s think about the questions emerged after studying the NBT paper.

Key findings of the NBT paper:

NgAgo/gDNA has met the basic requirements for an efficient gene editing system, and it has the ability to edit a mammalian genome efficiently in vivo in a “One Guide Faithful” way.

1. In the presence of a 5’p ssDNA guide(s), NgAgo acts as a nuclease, targets specific sequences and makes double strand break (DSB) in DNA

• in vitro, E. coli GST-NgAgo needs 2 guides (NBT paper Fig 1)

Note: This feature has been observed before for TtAgo (3), where TtAgo+1 guide can generate nick and TtAgo+2 guides can generate DSB on a target plasmid in vitro.

• in vitro, Mammalian FLAG-NgAgo-HA needs only 1 guide (NBT paper Fig 2, Fig 3a, 3b Supplementary Fig 1)

Note: Here is a breakthrough — NgAgo purified from mammalian cells only needs ONE guide for generating DSB on target DNA!

Note: This feature has not been reported for TtAgo.

Questions:

• Why is there a jump from 2 guides to 1 guide?
• Where are the data for purified NgAgo fusion proteins that are crucial for supporting the jump?

2. NgAgo could enter the nucleus of mammalian cells

• NLS-NgAgo is localized to the nucleus (NBT paper Supplementary Fig 4, Hela cells)

Questions:

• What was the detection method for NLS-NgAgo?
• What were the antibodies used, if any, e.g. antibody against NLS or against NgAgo, for NLS-NgAgo?

3. NgAgo/gDNA is able to edit mammalian genome

• Data presented in Fig 4, 5, Supplementary Fig 5,6,8 of the NBT paper

4. The efficiency of NgAgo/gDNA for gene editing in mammalian cells is comparable to that of CRISPR/Cas9

• Data presented in NBT paper Fig 4

5. NgAgo/gDNA has lower off target effect as compared to CRISPR/Cas9

• Data presented in NBT paper Supplementary Fig. 8

Note: There is no other quantitative data to support this point, such as data from deep sequencing.

Question: Why could other scientists not replicate the in vivo gene-editing function of NgAgo/gDNA, as was described in Fig 4, Fig 5 and other related figures in the NBT paper? (2)

With experimental details missing, one also has to wonder:

• Has anyone been able to replicate the in vitro DSB activity of NgAgo/gDNA as supported by Fig 1, Fig 2, Fig 3a, Fig 3b and Supplementary Fig 1 of the NBT paper?
• Has anyone been able to detect nuclear localization of NgAgo as supported by Supplementary Fig 4 of the NBT paper?

Section 4. Conclusions and Recommendations

After reviewing the NBT paper in detail, there is still no clear answer for why so many scientists could not replicate NgAgo for gene editing.

Instead, there are a few concerns arisen regarding the data, or the lack of, for NgAgo/gDNA as an efficient gene editing system:

1. NgAgo in generating DSB on target DNA in vitro and in vivo

• The jump from requiring “Two guides” to needing only “One guide”
• The lack of supporting protein data on purified NgAgo (GST-NgAgo, FLAG-NgAgo-HA) for in vitro DSB assays

3. Nuclear localization of NgAgo

• The lack of experimental details on the nuclear localization of NLS-NgAgo.

To conclude, here are the recommendations:

For those who still may be interested in testing NgAgo for gene editing

The following two approaches are recommended:

1. Follow the NBT paper from the beginning

• Get hold of GST-NgAgo, FLAG-NgAgo-HA and NLS-NgAgo vectors
• Try to replicate data for Fig 1, 2, 3 of the NBT paper
• If NgAgo could utilize one 5'p ssDNA guide to generate DSB on target DNA in vitro, proceed to in vivo work (Fig 4, 5 and Supplementary Fig 4 of the NBT paper).

2. Or, just test NgAgo in mammalian system

• Get hold of FLAG-NgAgo-HA and NLS-NgAgo vectors
• Try to replicate data for Fig 3a , 3b of the NBT paper
• If purified FLAG-NgAgo-HA could utilize one 5'p ssDNA guide to generate DSB on target DNA in vitro, proceed to in vivo work with NLS-NgAgo (Figure 4, 5 and Supplementary Figure 4 of the NBT paper).

But please be aware, other scientists have experienced difficulties in replicating Fig 4 and Fig 5 (2); therefore, when success could not be reached in the end, please do not be disappointed.

For authors of the NBT paper

They are the ones who have designed and conducted the experiments, and they are the possessors of raw data for the NBT paper.

To solve this “failure to replicate” mystery of NgAgo, it could be beneficial if the authors would be able to share the following information with the research community.

For NgAgo

No worries, NgAgo.

Nothing personal, just keep doing your job.

To edit or not, you will be fine; life goes on.

Note

Good science always enlightens a reader’s mind, while otherwise, it could be excruciating and exhausting.

If you have read as far as to the end of this blogpost, please find time to review these CRISPR/Cas9 papers. They will help refresh our memory and re-energize our brain cells.

Recommended readings for CRISPR/Cas9 technology

https://www.ncbi.nlm.nih.gov/pubmed/22745249

http://science.sciencemag.org/content/339/6121/819

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3712628/

https://elifesciences.org/content/2/e00471

Acknowledgement

Thank You for those who have helped review and edit the content of this blogpost; and Thank You for those who have kindly provided feedback.

Reference

1. NgAgo Nature Biotechnology article

http://www.nature.com/nbt/journal/v34/n7/full/nbt.3547.html

2. NgAgo “Failed to Replicate”

NgAgo

My experience with Natronobacterium gregoryi Argonaute (NgAgo)

Google Forums Round Up: First Impressions of NgAgo

Replications, ridicule and a recluse: the controversy over NgAgo gene-editing intensifies

New Genome Editing Tool Elicits Interest, Concern About Reproducibility

科学网-NgAgo你做出来了吗 - 蔡宇伽的博文

BioTechniques - To Edit or Not: The NgAgo Story

The Global Experimentation on NgAgo

Scholars call for probe into NgAgo claims - Global Times

Chinese Scientist's Rising Star Implodes

3.TtAgo Nature article

http://www.nature.com/nature/journal/v507/n7491/full/nature12971.html