Theory

Description.

Operation: PETase - only the tip of the iceberg

Over the past few decades the use of plastic has become exceedingly crucial in our daily lives. Plastic can be found almost everywhere, from common household objects to the undeniably beneficial medical applications. The reason why plastic has become that widely utilized is due to its many different properties. However, while all petro-based polymers show diverse attributes, there is one fundamental rule that binds them together - they need a considerably long time to decompose which can lead to devastating environmental consequences. PETase is a newly discovered enzyme able to break down highly resistant PET (polyethylene terephthalate) by hydrolyzation. It presents a promising alternative towards solving the ongoing plastic pollution problem. Still, the enzyme itself is not efficient enough to be applied in the field yet.
With our newly devised method D.I.V.E.R.T. we intend to enhance the catalytic activity of PETase in pursuit of unlocking its full potential. Since D.I.V.E.R.T. is a generalized concept, it can also be applied to other enzymes as well as binding proteins in any host organism. Presumably, it has the potential to revolutionize the face of directed evolution.

Design.

D.I.V.E.R.T. – directed in vivo evolution via reverse transcription

Abstract

In vivo continuous directed evolution offers significant advantages over classic in vitro methods as it drastically reduces the amount of time and actual lab work that needs to be invested. Most current approaches, however, are based on globally enhanced mutagenesis rates eventually leading to unwanted off-target mutations that interfere with the experiment. Here, we present the concept of a new continuous in vivo evolution strategy that allows complete spatial control of mutagenesis by cyclically using an RNA intermediate which finally replaces the original DNA cargo at the respective locus in the genome after it has been reverse-transcribed in an error prone way.

Introduction:

Finding and enhancing desired properties via directed evolution remains to be a key strategy in the engineering of biomolecules as rational design approaches are still limited due to deficient understanding of sequence to structure and function relations. The drawback of traditional in vitro laboratory evolution techniques, though, is that most steps in the procedure involve substantial human intervention and a single round of mutation, transformation, protein expression and selection usually requires several days to be completed¹ while for optimal results multiple rounds need to be performed².

Most of the in vivo methods that have been devised by scientists to overcome these difficulties so far either rely on global mutagenesis providing no spatial control, employ fusion proteins of DNA binding motifs and mutating enzymes like AID that only utilize a small section of the available sequence space or involve the assistance of phages imposing limits for the selection of phenotypes not well linkable to phage propagation³. Other approaches utilize engineered error-prone DNA polymerase I for replication of a cargo-carrying plasmid⁴. This way the mutational load on the genome is minimized, but regulatory elements on the plasmid are mutated just as likely as the coding sequence leading to differences in expression levels rather than protein activity. Also, since cells contain multiple copies of the gene of interest another pre-selection transformation step is necessary to finally generate the library that can be screened.

The iGEM project presented here allows for spatially fully controlled hypermutation in a means that is theoretically generalizable to prokaryotic as well as eukaryotic cells by creating an “artificial retroelement”. The idea takes the work of Crook et al.⁵ who inserted genetic cargo into a yeast retrotransposon to achieve continuous mutagenesis a step further by generating a system that undergoes the retrotransposon life cycle using an error prone reverse transcriptase (RT) that is encoded somewhere in trans rather than attaching the cargo to an already existing retroelement. This way, the concept can be extended to host species that lack suitable own retroelements like bacteria. Furthermore, the gene encoding the necessary reverse transcriptase is not mutated and regulatory sequences can be protected by placing them outside of the mutation cassette.

Detailed project description:

General:

One D.I.V.E.R.T. cycle comprises transcription, reverse transcription and reintegration of the gene of interest into the genome replacing the original copy (Figure 1). Since reverse transcriptases generally show poor fidelity mutations accumulate within the cargo while cycling. This overall scheme serves as an underlying principle for more elaborate arrangements described below. The individual approaches mainly vary in primer usage as well as the recombination mechanism applied. To find the optimal setup we plan on implementing all of them in E. coli and S. cerevisiae during the lab work phase starting in July.

Figure 1: General scheme of D.I.V.E.R.T. RNA is transparent.

Host difference considerations:

The following measures are taken in respect to the differences in mRNA processing between yeast and bacteria:
• In S. cerevisiae all relevant proteins are fused to a NLS to facilitate transport into the nucleus.

• To avoid interference between RT and ribosomes in E. coli the gene cassettes used are designed in a way that the transcript that is subject to reverse transcription is generated from the non-template DNA strand. This leads to a somewhat more complicated bidirectional cassette design.

• The in vivo generation of short RNAs in eukaryotes imposes some difficulties that also arise when transcribing gRNAs required by Cas9 in the CRISPR/Cas9 technology (lack of inducible promotors for RNA Pol III, unfavorable processing of transcripts from RNA Pol II). These can be overcome by using self-cleaving ribozymes enclosing the actual functional small RNA⁶. To have temporary control over primer generation in yeast we also intend to use this technique.

Also, in E. coli all functions necessary for D.I.V.E.R.T. (RT, Flp, primers) are encoded on a plasmid in contrast to being integrated into the genome via homologous recombination in yeast.

Approach I – Flp-FRT mediated recombination:

Just like any other DNA polymerase RT relies on a double stranded priming segment to kick off DNA synthesis. In Approach I an exogenous RNA primer is transcribed in trans and binds to the cargo mRNA enabling reverse transcription of the first strand (Figure 2).

Figure 2: D.I.V.E.R.T. Approach I; RNA is represented transparently.

The mRNA template then is degraded by the RNase H activity of the reverse transcriptase before the next primer can bind to finish reverse transcription which leads to a double-stranded pre-reintegration intermediate containing eventual point mutations that have been acquired in the process. The newly generated dsDNA as well as the original copy on the chromosome contain FRT sequences resulting in recombination through Flp-recombinase.

Drawbacks of this approach are that it relies on many components to work (primer, RT, Flp) and that mutations can occur within the FRT sequence, potentially impeding recombination.

Approach II – homologous recombination:

Employing homologous recombination instead of the Flp-FRT system allows to cut down to two components that need to be expressed in the host aside from the cargo. This is only true for yeast, though, since proteins of the lambda Red system are required to facilitate homologous recombination in E. coli⁷. The dsDNA produced by reverse transcription, however, should display at least one 3’ overhang (depending on when in the process the RNA primers are degraded by RNase H, Figure 3) so that only the beta protein from phage lambda is absolutely necessary.

To facilitate recombination it might even be sufficient to work in a RecA⁺ strain hence relying solely on the cell’s endogenous HR system equivalent to what is done in yeast. Both cases require the genes of the exogenous factors (primers and RT) to be integrated into the genome since plasmids cannot be stably sustained in recombination positive E. coli strains. For the project we will just go with beta, though.

Figure 3: D.I.V.E.R.T. Approach II; the possible forms of the dsDNA prior recombination are depicted on the right. RNA is represented transparently.

An additional advantage of this strategy is that it is more robust against unwanted mutations compared to Approach I where propagation of one single SNP in the FRT sequence might disable efficient recombination while it takes a higher number of point mutations in the primer binding site to actually prevent binding.

Approach III – ssDNA mediated recombination and self-priming:

Recombination in yeast^8,9 as well as in E. coli¹⁰ can be promoted by ssDNA oligonucleotides and there is even research suggesting that Red mediated dsDNA homologous recombination in E. coli involves a fully single-stranded intermediate¹¹. This is exploited in Approach III where (like Approach II) recombination only depends on the lambda beta protein (in E. coli) or does not need any exogenous factor at all (in yeast). To further reduce the number of components needed and since RT only requires one priming event to generate ssDNA, the cargo features complementary sequences leading to a dumbbell structure that can act as a double stranded priming site at the mRNA’s 3’ end (Figure 4). Due to RNase H activity only the ssDNA remains and is ready for recombination.

Figure 4: D.I.V.E.R.T. Approach III; RNA is represented transparently.

This concept allows D.I.V.E.R.T. to only require one component (the reverse transcriptase) in addition to the cargo cassette but also holds some disadvantages: To facilitate effective recombination, for example, one must make sure that the resulting ssDNA is complementary to the lagging strand of DNA replication. Moreover, ssDNA mediated recombineering is usually done using small oligonucleotides up to 100 nt and there is hardly any literature on the efficiency of incorporating longer DNA strands. Lastly, the secondary structure at the 3’ end might interfere with the secondary structure involved in transcription termination in E. coli and in yeast mRNA processing of the 3’ end most probably acts as a competing process. Thus, the feasibility of Approach III in yeast is highly speculative. We plan on trying to form the dumbbell structure by adding a sufficient number of Ts at the 3’ end of the cargo that should interact with the poly(A)tail but are skeptical.

Applications and Implications:

Rapidly mutating a single gene, pathway, regulatory sequence or any other cargo within a stable genome by just setting up the experiment and walking away sounds too good to be true. By allowing just this, a system like D.I.V.E.R.T. could drastically reduce the time and labor needed for carrying out traditional directed evolution experiments and also come in handy in many other fields of synthetic biology. Moreover, in spite of reverse transcriptases not offering error rates as high as those provided by some in vitro mutagenesis methods (e.g. error-prone PCR), in vivo-based techniques nonetheless enable the generation of larger libraries due to limits in transformation efficiency inherent to in vitro procedures. After optimizing their experimental design Crook et al. reported a potential library size of 3.7 × 10⁻² per round and cell. Since yeast and especially E. coli can reach cell densities of above 10¹² L^-1 (yeast only in controlled fermentation) theoretically infinite library sizes are only a matter of time and medium volume.

However, there is still considerable potential for optimization. For instance, mutation rates could be increased by engineering RTs or the RNA polymerase used for cargo-transcription towards lower fidelity. Also, the possible occurrence of unwanted mutations in sequences necessary for cycling (e.g. the FRT sequence in Approach I or primer binding sites in Approach II), although not posing an immediate problem due to the relatively short length of those sequences (and the robustness in primer binding in respect to single point mutations), could still be eradicated in future applications.

To summarize, we strongly believe that current methods, although finding broad application (e.g. in enzyme or antibody engineering), by far do not exploit the full potential of directed evolution. The ideal technique would enable researchers to unrestrictedly mutate distinct sequences without having to leave the living system. On the path towards this ultimate goal something like D.I.V.E.R.T. can be the next step.