<-chapter_16|Chapter 16^table_of_contents|Table of Contents^chapter_18|Chapter 18->
Chapter 17. Newer tools for %%reverse genetics%%
It is useful to briefly describe some newer technologies for reverse genetics that give researchers a bigger and more advanced toolbox for genetic analysis. Most of these tools are developed for use in mice, but generally they can be adapted to different model genetic organisms with some technical tweaks.
===== Gene knock-ins =====
Gene knock-ins are a variation of the classic gene knockout we saw in [[chapter_16|Chap. 16]]. In this method, instead of replacing an endogenous gene with something like $neo^R$, we replace it with a modified version of the gene. For instance, instead of going through the complex transgene/knockout combination strategy described in Chapters [[chapter_15|15]]-[[chapter_16|16]], you could in principle simply replace the mouse globin genes with knock-ins of the human globin genes. (When the studies described in Chapters [[chapter_15|15]]-[[chapter_16|16]] were originally done, the knock-in technique had not yet been invented.) This greatly reduces the complexity of the experimental design of the globin gene project. Furthermore, the level of expression of transgenes is often difficult to control precisely because transgenes often use upstream regulatory sequences that are removed from their normal chromosomal context (the regulatory sequences used may be missing some important enhancer sequences, for instance). By comparison, a knock-in allele usually will be expressed at very similar levels as the wildtype allele, because all the normal enhancer sequences that regulate the endogenous locus are still there.
To understand how gene knock-ins work, we must first discuss a site-specific recombination system such as the Cre-$lox$ system from bacteriophage P1 (Fig. {{ref>Fig1}}). $loxP$ is a 34 bp DNA sequence from the genome of bacteriophage P1. Normally, 34 bp is not long enough to drive homologous recombination in mice. However, a bacteriophage P1 enzyme called Cre (also called Cre recombinase) can cause DNA recombination (essentially crossing over) to occur between two different $loxP$ sites. This results in the sequences between the $loxP$ sites being deleted, leaving behind a single $loxP$ site as a recombination scar.
{{ :cre_lox_model.png?400 |}}
The Cre-$lox$ system for site-specific DNA recombination. Middle panel: Cre recombinase catalyzes recombination between two $loxP$ sites. If the two $loxP$ sites are facing in the same direction, the end product of this recombination is that the DNA in between is deleted as a circle, while leaving behind a single $loxP$ site as a recombination scar. If the $loxP$ sites face each other, then the sequence in between them becomes inverted (left panel). Finally, if the $loxP$ sites are in trans, it creates a translocation (right panel). Source: [[https://blog.addgene.org/plasmids-101-cre-lox|AddGene]]. Used with permission for non-commercial educational purposes.
To generate a gene knock-in, you would usually follow the normal gene knockout experimental approach, with a few modifications. For example, let's say you want to modify a codon in exon 2 of a gene (Fig. {{ref>Fig2}}). Here is how you might design your targeting construct:
* You would clone (probably using PCR) DNA sequences from exon 1, intron 1, and exon 2.
* You would insert $neo^R$ into intron 1, flanked by $loxP$ sites on either side.
* You would engineer the mutation you want into the appropriate sequence in exon 2.
* $tk^{HSV}$ would go after exon 2.
This targeting construct can be used to generate a "knockout", but since $neo^R$ is located in an intron, it doesn't really disrupt the gene at all. This targeting construct will integrate via homologous recombination using sequences from exon 1 and exon 2 in the targeting construct, thereby bringing in your engineered mutation to the correct chromosomal location in the ES cells. Once you have targeted ES cells, you can remove the $neo^R$ gene from the intron by expressing Cre in the ES cells before creating chimeric mice, leaving behind a $loxP$ scar in the intron. This is usually done with a separate transgene that expresses Cre.
{{ :knock_in.jpg?400 |}}
Gene knock-in strategy. The blue triangles represent $loxP$ sites. See text for details. Credit: M. Chao.
===== Conditional knockouts =====
Conditional knockouts are a variation of the gene knock-in method (Fig. {{ref>Fig3}}). A conditional knockout is a gene knockout that only occurs under certain conditions, such as in specific cell types or at specific developmental time points. To create a conditional knockout, you would use the gene knock-in method above to add two $loxP$ sites in two different introns of a gene you wanted to knock out. Since the $loxP$ sites are in introns, they typically will not affect the function of the gene; this allele usually will be indistinguishable from wildtype. If you can somehow express Cre in these cells, then Cre will cause recombination to occur between the two $loxP$ sites, which will delete all the exons in between. This usually results in a frameshift mutation, and frameshifts usually result in a null allele. In conditional knockout experiments, Cre is usually expressed as a transgene, and there are other tricks that can be done to express Cre only in specific cells, such as using [[chapter_17#Viral_vectors|viral vectors]], or to only express Cre under the control of an inducible promoter that depends on an external inducer, or some combination thereof.
{{ :conditional_ko.jpg?400 |}}
Conditional knockout strategy. The blue triangles represent $loxP$ sites, which are engineered into the chromosome using the gene knock-in strategy described in Chap. 17.1. Cre is usually expressed from a separate transgene (not shown). Credit: M. Chao.
Conditional knockouts are useful for studying genes that are expressed in multiple tissues, but you only wish to study the effect in one kind of tissue. For instance, let's say you are studying gene $X$, which is expressed in both brain and liver tissue. When you knock out gene $X$ using the standard approach, you find that the mice die as embryos due to a brain defect during embryogenesis. However, you are interested in the role of gene $X$ in the livers of adult mice. You can generate a conditional knockout of gene $X$ where it is only knocked out in adult livers but remains functional in embryonic brains so that they can be born and develop into adults that you can then study for liver function. One way to achieve this is to find a gene that is only expressed in adult livers, clone the upstream regulatory sequences that regulate expression in adult liver cells, and generate a transgene that expresses Cre under the control of this regulatory sequence. You can then generate a floxed allele (an allele with flanking $loxP$ sites) of gene $X$ and breed mice that carry this allele together with your liver-specific Cre transgene.
===== CRISPR/Cas9 =====
As we learned in [[chapter_16|Chap. 16]], homologous recombination occurs only at low frequencies in mammalian cells – this is why it was necessary to develop technologies to culture ES cells in vitro, so that we could examine large numbers of ES cells to look for rare events. It turns out that the frequency of homologous recombination increases enormously if there is a double-stranded break in the DNA near the site you wish to have homologous recombination. CRISPR/Cas9 (Fig. {{ref>Fig4}}) is an enzyme usually isolated from the bacterium //Streptococcus pyogenes// that can be used to cleave dsDNA to generate a double-stranded break. However, unlike restriction enzymes that only cut short defined palindromic sequences ([[chapter_09|Chap. 9]]), it is possible to modify CRISPR/Cas9 such that we can essentially make it cut almost any unique DNA sequence we wish. This is possible because CRISPR/Cas9 uses a guide RNA (gRNA) as part of its enzyme structure and the gRNA sequence is what guides it to its target DNA through RNA/DNA base pairing. It is relatively easy to modify this RNA sequence so that CRISPR/Cas9 will target the DNA sequence of a specific gene you are interested in. When dsDNA is cleaved in live cells, repair mechanisms will rejoin the cleaved DNA; however, errors often occur during this process, and indel mutations usually occur and therefore frameshifts, which usually creates null alleles. If a targeting construct is also present, homologous recombination will occur at a very high frequency; this allows for creation of knock-in alleles without necessarily using the Cre-$lox$-based strategy described above.
{{ :grna-cas9_w_bkgrnd.png?400 |}}
CRISPR/Cas9 is an enzyme that generates double-stranded breaks in target DNA sequences that match its guide RNA (gRNA). One requirement of CRISPR/Cas9 is that there must be a trinucleotide DNA sequence called a PAM sequence next to the cleavage site as well. Source: [[https://commons.wikimedia.org/wiki/File:GRNA-Cas9.png|Wikimedia]]. Credit: Marius Walter. Licensing: [[https://creativecommons.org/licenses/by-sa/4.0/deed.en|CC BY-SA 4.0]].
Using CRISPR/Cas9, it is possible to create knockouts and knock-ins in other model organisms such as //C. elegans// and zebrafish where culturing ES cells is not possible, and it is even possible to do it directly in somatic tissue without having to generate knockout/knock-in mice. CRISPR/Cas9-generated dsDNA breaks make homologous recombination so efficient that it has largely replaced the traditional method of generating knock-ins and knockouts using the double selection method. CRISPR/Cas9 can even be used for gene therapy in humans. If there is a disease allele in a patient, CRISPR/Cas9 can potentially be used to modify it so that it no longer has deleterious effects.
===== Viral vectors =====
Some viruses can be modified to deliver genetic material of any desired sequence into cells. Traditionally, viral vectors such as adeno-associated viruses (AAVs) or lentiviruses (a type of retrovirus) have been used to deliver modified genes into cultured cells in an analogous approach as generating transgenes. These types of viral vectors can also be used to deliver transgenes into live animals without using microinjection as described in [[chapter_15|Chap. 15]]. A modern experimental approach to knock out gene function that is much faster than generating knockouts or knock-ins with ES cells is to use a viral vector to deliver genes coding for a modified CRISPR/Cas9 enzyme directly into live tissue in an animal. This allows you to analyze loss of function of a gene in live tissue in a live animal without having to go through the trouble of creating a knockout mouse (although depending on your needs, a knockout mouse can also be very valuable).
Let's look at AAVs a bit more closely (Fig. {{ref>Fig5}}).
{{ :aav_structure.jpg?400 |}}
Structure of the viral capsid and genome of adeno associated virus (AAV). The top right panel shows the natural ssDNA genome of wildtype AAV. The viral genome codes for three different proteins (Rep, Cap, and AAP) using three different promoters (p5, p19, p40). ITR, inverted terminal repeat; pA, poly-A signal (related to terminating transcription and adding a poly-A tail to the mRNA). Recombinant AAVs (rAAV; bottom right panel) can be created where the viral genes have been replaced with a transgene (pink) and a promoter for the host cell you want to express the transgene in (green arrow). The ITRs are retained so that the rAAV genome can be replicated and packaged into a capsid inside a host cell (see Fig. {{ref>Fig6}}). Source: Marino and Holt (2022) Front. Neurol. 13 https://doi.org/10.3389/fneur.2022.870799. Licensing: [[http://creativecommons.org/licenses/by/4.0/|CC BY 4.0]].
Naturally occurring AAVs are small nonpathogenic parvoviruses. The AAV genome is just under 5 kbp long and is made of ssDNA instead of the more typical dsDNA. The ends of the ssDNA genome form structures called inverted terminal repeats (ITRs; Fig. {{ref>Fig5}}) that are required for replicating and packaging viral ssDNAs into new viral capsids. AAVs cannot successfully replicate in cells by themselves because they depend on proteins provided by another kind of virus (usually an adenovirus or herpesvirus) to produce new AAV viral particles after infection. For instance, three proteins from adenovirus, E2A, E4, and VA, are required for AAV replication.
To produce recombinant AAVs (rAAVs; Fig. {{ref>Fig6}}) as a gene delivery vector for research or therapeutic purposes, you would need to construct three separate plasmids in vitro:
* A plasmid that expresses the Rep and Cap proteins that are normally encoded by the AAV genome (purple in Fig. {{ref>Fig6}});
* A second plasmid that contains your transgene of interest you are trying to deliver into cells (red in Fig. {{ref>Fig6}});
* A third plasmid that expresses the essential E2A, E4, and VA genes from adenovirus (green in Fig. {{ref>Fig6}})
You would then introduce all three plasmids into cultured cells (such as HEK293 cells, a human embryonic kidney cell line). The purple and green plasmids, together with the HEK293 host cell genome, provide all the proteins needed to replicate and package newly replicated viral ssDNAs. However, only DNA with ITRs will be packaged. This means that the plasmids containing AAV and adenoviruses will not be packaged (they won't even be replicated) - only the ssDNA with the ITRs, and therefore your transgene, will be packaged into new viral capsids. You can harvest the rAAVs from the media of the HEK293 cell culture.
{{ :aav_pacakaging.jpg?400 |}}
Strategy for producing recombinant AAVs (rAAVs) for research or gene therapy. See text for details. Source: Wang et al. (2024) Sig. Transduct. Target. Ther. 78 https://doi.org/10.1038/s41392-024-01780-w. Licensing: [[http://creativecommons.org/licenses/by/4.0/|CC BY 4.0]].
The rAAVs can then be used to deliver your gene payload into target cells through infection. Different serotypes (or variations) of AAV capsid proteins can be used to target your rAAVs to different tissue types (Table {{ref>Tab1}}). These rAAVs cannot produce new rAAV capsids after infection of target tissue because they do not encode the essential AAV genes Rep, Cap, and AAP. In principle, this makes rAAVs ideal for a vehicle for gene therapy. One drawback of AAVs is that there is a size limit to their DNA capacity - around 5 kbp is the maximum payload size. There are some tricks that can be used to get around this size limit, but it can be inconvenient if the transgene you wish to deliver into cells is greater than 5 kbp in length.
Some example of AAV serotypes (variants) with different tissue targeting specificity (tropism). Adapted from Naso et al. (2017) BioDrugs 31(4): 317-334 https://doi.org/10.1007%2Fs40259-017-0234-5. Licensing: [[http://creativecommons.org/licenses/by-nc/4.0/|CC BY-NC-4.0]].