DNA plasmids are foundational tools in molecular biology and are increasingly being used to generate lifesaving genetic medicines. These circular DNA molecules, naturally found in bacteria, are leveraged by researchers and physicians to deliver genes into foreign cells.
In the research world, we often borrow DNA plasmids from neighboring labs rather than designing them from scratch each time.
Such sharing reduces duplicated efforts, but it also means that the person using the plasmid is often not the person who designed the plasmid. So how do you know if the plasmid you borrowed is any good? And, how do you know if a once good plasmid has gone bad over time?
A new study found serious errors in up to 50% of DNA plasmids submitted by academic and industrial labs. The errors were particularly frequent in plasmids designed for gene therapy treatments. These findings highlight the need to carefully examine and verify plasmids before experimental and therapeutic use.
You may have heard the expression before about borrowing a cup of sugar from your neighbor. But in this case, it would be as if your neighbor gave you salt instead of sugar, and you forgot to verify before adding it into your recipe. Those will be some salty cookies!
With the use of plasmid sharing resources on the rise, and the increased use of plasmids to make genetic medicines, it is more important than ever to make sure that the plasmids you make, and use, are accurate. Thankfully, with next-generation sequencing techniques it is also easier than ever to verify your plasmids.
In this article:
Consequences of Wrong Plasmids
Analyzing Thousands of Plasmids
Mutations in Gene Delivery Plasmids
Consequences of Wrong Plasmids
Using faulty tools, such as incorrect plasmids, is one reason that billions of dollars are wasted on research that results in incorrect and irreproducible findings (Belluz, 2024; Ioannidis et al, 2014). Additionally, these errors lead to months and years of lost research time and frustratedly chasing false results.
As dramatic as that might sound, the stakes are even higher with plasmids that are used to generate genetic medicines. But first we need to cover a little background.
Adeno-associated viruses (AAV) are used for gene delivery therapeutics in diseases such as hemophilia and Duchenne muscular dystrophy (Crisafulli et al, 2020; Samelson-Jones & George, 2023). What’s really exciting is that many more AAV treatments are in development for a wide range of diseases (Wang et al, 2024).
In AAV treatments, the viral icosahedron – the outside protein armor of the virus – is reconstituted from recombinant viral proteins, with a gene packaged inside (Figure 1). This artificial virus will then deliver the gene to diseased cells to treat a disease. Since this gene-delivery virus is reconstituted, or built, from individual parts, there is no risk of it replicating in human cells like a real virus would. However, as a side note, our immune system doesn’t always know the difference, so devising ways to hide these gene-delivery viruses from the immune system is still an active area of AAV optimization (Liu et al, 2024; Long et al, 2024).
Figure 1. Adeno-associated virus (green) with DNA (purple) inside.
You may be wondering – what do DNA plasmids have to do with this? Well, plasmids are designed to deliver genes into the artificial virus. To do this, the gene is flanked by regions called inverted terminal repeats, or ITRs for short. ITRs are important for getting the gene inside the viral icosahedron while it’s assembling (Figure 2)(Earley et al, 2020).
ITRs are very rich in guanosine (G) and cytosine (C) DNA nucleotides. These GC-rich regions have distinct structures from canonical DNA base pairing, which is important for their viral packaging function, but also makes them highly mutation prone (Wang et al, 2024). When DNA polymerase is making a copy of ITRs, it gets stuck in the structured regions and makes mistakes resulting in mutations in the DNA sequence.
As we’ll discuss later in this article, the inherent structure of ITRs makes them especially prone to develop mutations over time. These ITR mutations also drastically reduce the efficiency of recombinant AAV production and reduce the specificity of DNA loading (Savy et al, 2017; Wilmott et al, 2019). In practice this means that it is more expensive to produce enough AAVs to treat patients, if it’s even feasible at all, and that an increased proportion of AAVs will have the wrong DNA sequence instead of the therapeutic gene – further reducing the AAV treatment’s potency.
Figure 2. The inverted terminal repeats (ITR) are important for DNA (purple) recruitment into adeno-associated viruses (green).
Analyzing Thousands of Plasmids
With that background in mind, I think you can see why the stakes are high in making sure that plasmids contain correct and functional DNA sequences.
The company VectorBuilder is a research service provider that clones plasmids, produces viral vectors, and performs experiments for clients. Over the years their clients have sent thousands of different plasmids to VectorBuilder for further modification. Dr. Bruce Lahn, Chief Scientist at VectorBuilder, notes that they were “both shocked and depressed to see [plasmid] errors cropping up over and over with such high frequency” (email communication, September 2024). Overtime, they realized they should thoroughly analyze these errors and communicate their findings to the broader research community.
In their recently published preprint, VectorBuilder looked at 1,132 plasmids provided to them from researchers from across the world (Bai et al, 2024). Out of the 1,132 plasmids analyzed, 852 of them had restriction enzyme sites flanking the gene of interest according to plasmid maps provided by the submitting labs. Restriction enzymes (RE) are molecular scissors used to cut specific DNA sequences.
They used restriction enzyme digestion to examine the overall structure of the plasmid. Basically, this determined whether the annotated restriction enzyme sites were there, and whether a gene of roughly the right size was in between these restriction sites (Figure 3).
Figure 3. Cutting
a gene of interest out of a plasmid. Left, restriction enzymes (scissors) cut
the plasmid on each side of the gene of interest (Gene X). Right, agarose gel
of the DNA plasmid without (-) and with (+) restriction enzyme treatment.
Restriction enzymes cut the gene of interest out of the plasmid resulting in a
smaller DNA band that runs faster on an agarose gel. In the second lane with
restriction enzymes the cut-out gene is too small so this plasmid is wrong. In
the third lane with restriction enzymes the plasmid doesn’t cut so it is wrong.
They found that 128 of the 852 plasmids with reported
restriction enzyme sites, about 15%, either didn’t cut at all or yielded DNA
fragments with sizes inconsistent with the reported RE site locations (Figure 3).
This indicated that either the plasmid was just flat out the wrong plasmid or contained point mutations in the RE sites rendering them nonfunctional for digestion.
To better determine what the issues with these plasmids might be – they sequenced almost 400 plasmids and found that about 32% of them had sequence errors including point mutations, deletions, and insertions. Many plasmids even had multiple types of errors.
Mutations in Gene Delivery Plasmids
While the results so far weren’t great, VectorBuilder found even more errors in plasmids designed for gene delivery in human therapeutics.
In their studies, VectorBuilder found that about 40% of the AAV plasmids they sequenced had mutations in at least one of the ITRs, with the ITR upstream of the delivered gene being much more frequently mutated.
Taking all of the different plasmid types together, their findings suggest that a substantial proportion, approximately 40 to 50%, of plasmids made in academic and industry settings contain mutations that may affect their function. The authors note that since their samples are self-selected for researchers who were trying to make sure their plasmids are correct, the actual frequency of mutations may be even a bit higher than they found.
Historical Perspective
Not too long ago when you requested a plasmid from another lab or a colleague, all you would get to guide you was something like the drawing in Figure 4. So, assuming the drawing is correct, you had at least some information about the plasmid you were working with. For example, you would know:
- The gene being expressed
- The antibiotic resistance being used for expression
- Maybe the promoter or restriction enzyme sites used to clone the gene of interest into the plasmid
Figure 4. A hypothetical plasmid containing Gene X and Ampicillin resistance (AmpR).
But this bare bones depiction of a plasmid kind of leaves you wanting to know more about it, right? For example:
- What is the exact sequence of the gene being expressed?
- Which gene isoform is being used?
- Are there any mutations in the gene of interest?
- Are there any mutations in the plasmid backbone, especially mutations that would render the plasmid nonfunctional?
- What restriction enzymes were used to clone the plasmid?
As you can tell from VectorBuilder’s study, such a sparse depiction of the plasmid, as in Figure 4, doesn’t inform about the plasmid at a granular level. Importantly, this missing detailed information may be very critical for the plasmid’s intended use.
How to Verify Your Plasmids
One easy way to examine your plasmid is to cut out the inserted gene using the flanking restriction enzyme sites (Figure 5). Recall this was VectorBuilder’s first check in their study. This method will inform you about the overall structure of the plasmid, and let you know if the RE sites are intact (Figure 1).
However, there are many more additional features about the plasmid that you will know nothing about if you only perform a RE digestion (Figure 5). By the way, check out this article if you want more in depth information about key plasmid features.
Figure 5. Key plasmid features that are important for plasmid function and to further modify the plasmid.
To learn more about your plasmid, it is always good to sequence it so you know exactly what you are dealing with at base pair resolution. Sanger sequencing, which is what VectorBuilder used in their study, is great for sequencing small to medium size genes and may even include part of the flanking regions of the plasmid. However, you can instead use a next-generation sequencing technique, such as nanopore sequencing, to get the sequence of the entire plasmid. Then you would know your plasmid in fine detail in its entirety.
If you want more information about how restriction enzyme digestion and different sequencing technologies are used to analyze DNA plasmids, this is article for you!
While this all might sound a bit tedious, verifying your plasmid up front can save you months or years on your research project, not to mention the headache of having to redo everything with the correct plasmid!
It’s a little scary to think about plasmids being so frequently wrong – especially those that are intended for therapeutic use. If you needed any encouragement to verify your plasmids before you start working with them - hopefully this research is the little boost you need to do so. At the very least, make sure you verify your plasmids before submitting them to VectorBuilder so that you can be on the right side of the statistics in their next study!
