1. Bibliographic Information

1.1. Title

Directed evolution of an orthogonal transcription engine for programmable gene expression in eukaryotes

1.2. Authors

Shaunak Kar, Elizabeth C. Gardner, Kamyab Javanmardi, Daniel R. Boutz, Raghav Shroff, Andrew P. Horon, Thomas H. Segall-Shapiro, Andrew D. Ellington, and Jimmy Gollihar. The authors represent affiliations from the Institute and Department of Pathology and Genomic Medicine at Houston Methodist Hospital, the Department of Bioengineering at Rice University, and the Department of Molecular Biosciences and the Center for Systems and Synthetic Biology at The University of Texas at Austin. This multidisciplinary authorship indicates a strong background in genetics, molecular biology, biomolecular engineering, and synthetic biology. Shaunak Kar and Elizabeth C. Gardner contributed equally to the work, and Jimmy Gollihar is the lead contact.

1.3. Journal/Conference

The paper was published in iScience, a multidisciplinary journal from Cell Press that publishes original research in the life, physical, and earth sciences. iScience is a reputable journal known for publishing significant advances across scientific disciplines, often at the intersection of basic research and applied science.

1.4. Publication Year

2024

1.5. Abstract

The paper describes the engineering and evolution of a novel orthogonal transcription engine for eukaryotes. The researchers fused the T7 RNA polymerase (RNAP) with a capping enzyme from African swine fever virus (NP868R) to enable co-transcriptional RNA capping, a crucial modification for functional mRNA in eukaryotes. Through directed evolution in Saccharomyces cerevisiae, they successfully enhanced the activity of this T7 polymerase fusion enzyme by approximately two orders of magnitude. The engineered system demonstrates programmable and orthogonal gene expression in both yeast and mammalian cells, offering tunable and scalable control for various synthetic biology applications.

1.6. Original Source Link

/files/papers/69215d363f69abbc8fd2082e/paper.pdf The publication status is officially published, as indicated by the provided publication date and journal information.

2. Executive Summary

2.1. Background & Motivation

The core problem the paper aims to solve is the challenge of achieving truly orthogonal and programmable gene expression in eukaryotic systems. While the T7 RNA polymerase (RNAP) system has been a cornerstone for orthogonal gene expression in prokaryotes due to its high specificity and independence from host machinery, its utility in eukaryotes has been severely limited. This limitation stems primarily from the absence of essential post-transcriptional modifications, particularly the 5' methyl guanosine cap (5' cap), on T7 RNAP-derived transcripts. Functional eukaryotic mRNAs require this 5' cap for efficient translation, nuclear export, and stability.

This problem is important because engineering complex genetic circuits in eukaryotes is difficult. Native eukaryotic gene expression relies on intricate host machinery (e.g., RNA polymerase II (Pol II), transcription factors, multi-component holoenzymes) and large, complex promoters, making predictability and orthogonality challenging. The ability to express genes independently of the host's transcriptional machinery would enable more robust, predictable, and portable synthetic biology applications across diverse eukaryotic hosts, from industrial yeast strains to mammalian cell lines for therapeutics.

The paper's entry point and innovative idea revolve around overcoming the capping limitation by creating a fusion enzyme. They combine the T7 RNAP with a viral capping enzyme (NP868R from African swine fever virus) that can form the 5' cap independently of host machinery. Critically, they then employ directed evolution in a yeast model to significantly enhance the activity of this fusion enzyme, addressing the low expression capacity observed in previous attempts with similar fusion enzymes.

2.2. Main Contributions / Findings

The paper makes several primary contributions:

• Engineered capping-T7 Fusion Enzyme: The development of a fusion enzyme, NPT7, combining T7 RNAP with the single-subunit capping enzyme NP868R. This enzyme is capable of co-transcriptionally capping RNA in eukaryotes, enabling the production of functional, capped mRNA outside of the host's native Pol II machinery.

• Directed Evolution for Enhanced Activity: The successful application of directed evolution in Saccharomyces cerevisiae to significantly improve the NPT7 fusion enzyme's activity. This process yielded variants ($v433$ and $v443$) that exhibited nearly two orders of magnitude (100-fold) higher protein expression compared to the wild-type fusion enzyme.

• Programmable and Orthogonal Gene Expression: Demonstration that these evolved NPT7 variants enable predictable and tunable control of gene expression in both yeast and mammalian cells. This includes the ability to modulate expression levels using known T7 promoter variants and to achieve multiplexed, independently controlled gene expression within a single system. The system maintains orthogonality to host factors.

• Cross-Kingdom Portability: Evidence that the yeast-evolved variants show enhanced performance (3-4 fold higher reporter expression) in mammalian HEK293T cells, suggesting broad applicability across diverse eukaryotic systems.

• Tunable and Scalable Control for Synthetic Biology: The establishment of a robust, orthogonal gene regulatory system that offers precise control over gene expression, enhancing the versatility and efficiency of synthetic biology applications in eukaryotes, such as refactoring metabolic pathways or improving mRNA therapeutic production.

  These findings directly address the challenge of orthogonal gene expression in eukaryotes by providing a highly active and controllable system that bypasses limitations of host machinery, thereby solving issues of predictability, tunability, and portability that previously hindered complex eukaryotic circuit design.

3. Prerequisite Knowledge & Related Work

3.1. Foundational Concepts

To fully understand this paper, a reader should be familiar with the following concepts:

• T7 RNA Polymerase (RNAP): A single-subunit enzyme from bacteriophage T7 that transcribes DNA into RNA. It is highly specific to a short 17 base pair (bp) promoter sequence and is known for its high transcription rate and orthogonality (i.e., it operates independently of the host's native transcriptional machinery). It has been widely used in prokaryotes for recombinant protein production and synthetic biology applications.
• Eukaryotic Gene Expression: The process by which genetic information from DNA is used to synthesize functional proteins in eukaryotes. This involves several key steps:
  • Transcription: DNA is transcribed into pre-mRNA by RNA polymerase II (Pol II).
  • Pre-mRNA Processing: The pre-mRNA undergoes crucial modifications:
    • 5' Capping: The addition of a 7-methylguanosine (m7G) cap to the 5' end of the mRNA. This cap is essential for mRNA stability, nuclear export, splicing, and efficient translation initiation. It distinguishes functional mRNA from other RNAs.
    • Splicing: Removal of non-coding introns and ligation of coding exons.
    • 3' Polyadenylation: The addition of a poly(A) tail (a long sequence of adenosine residues) to the 3' end of the mRNA, which contributes to mRNA stability, nuclear export, and translation.
  • Translation: Ribosomes use the mRNA template to synthesize proteins.
• Orthogonality: In synthetic biology, orthogonality refers to the ability of a genetic system (e.g., an RNA polymerase and its promoter) to function independently of the host cell's native regulatory pathways. This isolation allows for precise and predictable control without unintended interactions or interference from the host.
• Nuclear Localization Signal (NLS): A short amino acid sequence that tags proteins for transport from the cytoplasm into the cell nucleus by nuclear import receptors. In this paper, NPT7 is fused with an NLS to ensure it functions in the nucleus where the target DNA is located.
• Untranslated Regions (UTRs): Regions of an mRNA molecule that are not translated into protein. The 5' UTR is located upstream of the start codon, and the 3' UTR is downstream of the stop codon. Both play roles in regulating mRNA stability, localization, and translation efficiency.
• Directed Evolution: A laboratory method used to engineer proteins or nucleic acids with desired properties by applying cycles of mutagenesis, selection (or screening), and amplification. It mimics natural selection in a controlled environment, allowing for rapid optimization of biological molecules. In this paper, it's used to improve the activity of the NPT7 fusion enzyme.
• Fluorescence-Activated Cell Sorting (FACS): A specialized type of flow cytometry that allows for the physical separation of cells based on their fluorescent properties. Cells expressing a fluorescent reporter (like ZsGreen in this study) above a certain threshold can be isolated, enriching for variants with enhanced activity.
• Saccharomyces cerevisiae (Yeast): A well-characterized eukaryotic model organism commonly used in genetic and synthetic biology research due to its ease of manipulation, rapid growth, and conserved eukaryotic cellular processes.
• HEK293T Cells: A human embryonic kidney cell line commonly used in molecular biology and biotechnology due to its ease of culture and transfection, often used for recombinant protein production and gene expression studies in mammalian systems.

3.2. Previous Works

The paper builds upon a significant body of prior research attempting to harness T7 RNAP in eukaryotes and address the capping challenge:

• T7 RNAP in Prokaryotes: As mentioned, T7 RNAP has been a "cornerstone" in biotechnology for decades, particularly in Escherichia coli (Refs. 1, 2). Its high specificity to its 17 bp promoter and independence from host machinery make it ideal for high-yield RNA production and complex genetic circuitry (Refs. 3-5). Engineered derivatives have been used for controllers, resource allocators, and logic gates (Refs. 6-11).
• Challenges of T7 RNAP in Eukaryotes: Porting T7 RNAP to eukaryotes is challenging because functional eukaryotic mRNAs require post-transcriptional modifications, especially the 5' cap and poly(A) tail (Refs. 19-21). The absence of these on T7 RNAP transcripts severely limits protein expression in eukaryotes. While poly(A) addition can be encoded in DNA (Ref. 25), the 5' cap is problematic.
• Prior Capping Attempts:
  • Host Capping Machinery: Efforts to recruit host capping machinery by fusing Pol II-derived signaling domains to T7 RNAP were largely unsuccessful in generating appreciable protein expression (Refs. 26, 27). This highlights the difficulty of integrating T7 RNAP into the complex, coordinated host pathway.
  • Viral Capping Enzymes: Viral capping enzymes offer an alternative because they can catalyze cap formation independently of host machinery. The vaccinia capping enzyme (VCE) was noted for its role in cytoplasmic capping (Refs. 32-35). Co-expression of T7 RNAP and VCE has been shown to produce capped transcripts and protein in mammalian cell cytoplasm (Refs. 30, 31, 36). However, levels of capped transcripts were often low (Ref. 36), possibly due to uncoordinated activity.
  • Fusion Enzymes (T7 RNAP + Capping Enzyme): More recently, fusion enzymes combining T7 RNAP and VCE were shown to mediate protein expression in mammalian cells, offering enhanced expression compared to unlinked enzymes (Ref. 28). The capping enzyme from African swine fever virus (NP868R) was identified as a superior fusion partner, showing the highest protein expression in HEK293T cells (Ref. 28). Despite these advances, the protein expression levels were still significantly lower than native nuclear expression, indicating a need for further optimization (Ref. 28). The NP868R capping enzyme was specifically highlighted for its role in reovirus rescue and protein expression (Ref. 37).
• Orthogonal T7 RNAP Variants: Previous work has engineered T7 RNAP variants to recognize distinct mutant T7 promoters with minimal cross-reactivity (Ref. 6). This provides a toolkit for multiplexed and orthogonal gene expression, but these variants had not been successfully ported into eukaryotic hosts with high efficiency prior to this study.

3.3. Technological Evolution

The evolution of orthogonal gene expression technologies has moved from simple bacterial systems to increasingly sophisticated eukaryotic platforms:

1. Early T7 Systems (Prokaryotes): The journey began with the discovery and widespread adoption of T7 RNAP in E. coli in the 1980s. This provided a simple, powerful way to express foreign genes at high levels, decoupled from bacterial regulation.

2. Initial Eukaryotic Adaptation Attempts (Co-expression): Researchers recognized the potential of T7 RNAP for eukaryotes but immediately faced the challenge of eukaryotic mRNA processing, particularly the 5' cap. Initial attempts involved co-expressing T7 RNAP with viral capping enzymes like VCE, or trying to recruit host machinery. These often yielded low efficiency.

3. Fusion Enzyme Development: The logical next step was to physically link the T7 RNAP and capping enzyme into a single protein. This ensured co-localization and potentially better coordinated activity. The identification of NP868R as a potent capping enzyme for fusion was a key step here. However, even these fusion enzymes still suffered from relatively low activity compared to native eukaryotic expression.

4. Directed Evolution for Optimization (This Paper): This paper represents a significant leap by employing directed evolution to drastically improve the activity of the fusion enzyme. This approach leverages the power of high-throughput screening in a eukaryotic model (yeast) to rapidly identify highly efficient variants, effectively overcoming the inherent limitations of the initial fusion design.

5. Programmable and Multi-modal Systems: By integrating the highly active evolved variants with established T7 promoter libraries and demonstrating multiplexing, the paper pushes the technology towards truly programmable and complex circuit design in eukaryotes.

   This paper's work fits within this timeline by addressing the critical gap in fusion enzyme efficiency through directed evolution, making the concept of an orthogonal, capping-T7-based transcription engine viable for practical eukaryotic synthetic biology.

3.4. Differentiation Analysis

Compared to the main methods in related work, this paper's approach offers several core differences and innovations:

• Enhanced Activity via Directed Evolution: The most significant differentiation is the use of directed evolution to enhance the NPT7 fusion enzyme's activity by nearly two orders of magnitude. Previous fusion enzymes, while functional, showed much lower expression than native nuclear systems. This study systematically optimizes the enzyme itself, leading to a qualitative jump in performance.

• Yeast-Based Evolution for Broad Applicability: While previous NPT7 work primarily focused on mammalian cells (e.g., HEK293T), this study utilizes Saccharomyces cerevisiae as the primary platform for directed evolution. This is innovative because yeast offers easier genetic manipulation and high-throughput screening, and the resulting evolved enzymes show improved activity across kingdoms (yeast and mammalian cells), demonstrating broad portability.

• Demonstrated Programmability and Tunability in Eukaryotes: The paper rigorously demonstrates that the evolved NPT7 variants enable predictable and tunable gene expression using established T7 promoter libraries. This level of fine-grained control and modularity, orthogonal to host factors, was not as effectively achieved with previous, lower-activity fusion enzymes in eukaryotes.

• Multiplexed and Orthogonal Circuitry Validation: The study goes further to show that the system can support multiplexed gene expression with independent control over multiple targets simultaneously, and can incorporate previously characterized orthogonal T7 RNAP variants. This moves beyond simple protein expression towards complex genetic circuit design in eukaryotes.

• Direct Solution to 5' Capping Problem: While previous work identified the problem and suggested viral capping enzymes as a solution, this paper provides a highly optimized, single-protein solution that intrinsically handles both transcription and crucial 5' capping, making it a more robust and self-contained "orthogonal transcription engine."

  In essence, while the concept of T7 RNAP/capping enzyme fusions existed, this paper's innovation lies in using directed evolution to significantly boost their activity and then thoroughly validating their programmatic control and cross-kingdom utility, thereby transforming them from a proof-of-concept to a highly functional tool for eukaryotic synthetic biology.

4. Methodology

4.1. Principles

The core principle behind the methodology is to create an entirely orthogonal transcriptional system in eukaryotes that is independent of the host's native RNA polymerase II (Pol II) machinery and its associated factors. This is achieved by fusing the highly specific bacteriophage T7 RNA polymerase (RNAP) with a viral capping enzyme (NP868R) into a single protein, named NPT7. This fusion enzyme is designed to perform both transcription and the essential 5' methyl guanosine (5' cap) modification co-transcriptionally, thus producing functional mRNA in eukaryotic cells. The low initial activity of this fusion enzyme is then dramatically improved through a process of directed evolution, mimicking natural selection in the lab, to generate highly efficient variants. The system's orthogonality, tunability, and portability are subsequently validated across different eukaryotic hosts and genetic circuit designs.

4.2. Core Methodology In-depth (Layer by Layer)

The methodology can be broken down into several key stages: construction of the initial fusion enzyme, setting up the yeast screening platform, directed evolution, and characterization of the evolved variants in yeast and mammalian cells.

4.2.1. Construction of the Initial NPT7 Fusion Enzyme

The starting point was an existing concept of fusing a T7 RNAP with a viral capping enzyme. Here, the specific components were:

• T7 RNAP: The RNA polymerase from bacteriophage T7.

• NP868R: The capping enzyme from the African swine fever virus, previously identified as effective for fusion. These two components were fused together to create NPT7. To ensure the enzyme functions in the correct cellular compartment:

• Glycine-serine linker: A flexible linker was used between the two protein domains to allow for independent folding and function.

• N-terminal nuclear localization signal (NLS): An NLS was added to the fusion protein to direct it into the nucleus of eukaryotic cells, where the target DNA templates are located.

  The gene encoding this NPT7 fusion enzyme (with linker and NLS) was placed under the control of a galactose-responsive promoter. This allows for inducible expression of the NPT7 enzyme, where varying concentrations of galactose can control the amount of NPT7 produced. This entire cassette was integrated into the HO locus of the Saccharomyces cerevisiae (yeast) genome. The HO locus is a common site for stable genomic integration in yeast.

To test the activity of NPT7, a reporter construct was designed:

• ZsGreen: A gene encoding a fluorescent protein (ZsGreen) was used as the readout.
• T7 RNAP promoter: The ZsGreen gene was placed under the control of a T7 RNAP promoter, meaning its transcription would only be initiated by the NPT7 enzyme.
• Polyadenylation signal and T7 terminator: These sequences were included downstream of ZsGreen to ensure proper mRNA processing and termination in yeast. This reporter construct was encoded on a multi-copy yeast plasmid.

The activity of this initial setup was assessed by transforming the NPT7 and reporter constructs into the Saccharomyces cerevisiae strain BY4741. Upon induction with galactose, a ~2-fold increase in ZsGreen protein production was observed compared to controls (Figure 1B, 1C), confirming low but detectable activity.

To confirm that the observed protein expression was dependent on the capping activity of NP868R, a control experiment was performed:

• Capping-null mutant: A point mutation (K294A) was introduced into the NP868R domain of NPT7. This K294A mutation is known to disrupt cap formation, creating a "capping dead" version of the enzyme. Transformation and induction of this capping-null NPT7 mutant resulted in significantly reduced reporter expression (Figure S1B), confirming the critical role of the capping enzyme in the fusion.

4.2.2. Directed Evolution of NPT7

Recognizing the low activity of the wild-type NPT7, a directed evolution strategy was employed to enhance its function:

1. Mutagenesis: Error-prone PCR was used to generate a library of NPT7 variants. Error-prone PCR introduces random mutations into the DNA sequence, creating a diverse pool of enzyme variants.

2. Selection/Screening in Yeast:

   • The NPT7 variant library was transformed into Saccharomyces cerevisiae cells containing the ZsGreen fluorescent reporter construct.
   • Fluorescence-activated cell sorting (FACS): Cells were sorted based on their ZsGreen fluorescence intensity. The top 1% of cells exhibiting the highest ZsGreen expression were isolated. This selects for NPT7 variants with enhanced activity.

3. Amplification and Iteration: The selected cells were grown, and their NPT7 genes were extracted, amplified, and subjected to further rounds of error-prone PCR and FACS, mimicking an evolutionary process to progressively enrich for more active variants.

   After several rounds of selection, two highly active variants were isolated: $v433$ and $v443$. These variants showed nearly two orders of magnitude (approximately 100-fold) increase in ZsGreen signal relative to the wild-type NPT7 (Figure 1C).

4.2.3. Genotyping and Characterization of Evolved Variants

• Mutation Analysis: The genes encoding $v433$ and $v443$ were sequenced to identify the mutations responsible for the enhanced activity.
  • $v433$ contained 15 amino acid mutations (7 in NP868R and 8 in T7 RNAP).
  • $v443$ contained 10 amino acid mutations (2 in NP868R and 6 in T7 RNAP).
  • Convergence was observed at five specific residues (K9N, D279N, H1195R, R1202H, and H1667R) between the two variants (Figure 1D; Table S1).
• Confirmation of Capping Dependency: Capping-null versions of $v433$ and $v443$ were created (by introducing the K294A mutation in NP868R). These inactivated evolved variants showed significantly lower reporter expression than their functional counterparts, confirming that the increased activity still primarily relied on the capping function (Figure S1B). However, the capping-null evolved variants still showed ~10-fold higher activity than the wild-type capping-null NPT7, suggesting a low level of background capping activity or other non-capping-related improvements.
• Genomic Target Testing: The evolved NPT7 variants did not show significant activity on genomic targets in yeast (Figure S1C), suggesting that repressive chromatin conditions might inhibit their function on integrated genes. Therefore, subsequent experiments focused on high-copy plasmid reporters.

4.2.4. Tunable Control and Orthogonality Characterization in Yeast

• Titratable Expression (Enzyme Level): To demonstrate tunable control, a $ΔGal2$ yeast strain was used. This strain allows for titratable expression of genes under the galactose-responsive promoter by varying galactose concentrations. By controlling the amount of galactose, researchers achieved graded levels of reporter expression, demonstrating that the final gene expression can be regulated by tuning the level of the fusion enzyme (Figure 1F).
• Comparison to Host Promoters: The expression level of $v443$ was compared against endogenous host-regulated promoters. A ZsGreen reporter was placed under the control of the strong galactose-responsive promoter and tested in two contexts: integrated into the Leu locus (IV target) and on a multi-copy 2 micron plasmid (2u target). The reporter expression from $v443$ was found to be comparable to that driven by the endogenous Gal promoter (Figure 1H).
• Nuclear Activity Confirmation: Experiments were performed with NPT7 variants lacking the NLS tags. Since the reporter plasmid is known to be nuclear in yeast, removal of the NLS should reduce reporter levels if the enzyme's activity is indeed nuclear. A significant reduction in reporter expression was observed for non-NLS versions compared to NLS versions (Figure S2B), confirming the nuclear activity of the enzymes.
• Polyadenylation Signal Swapping: To simplify circuit design, different endogenous yeast polyadenylation (pA) signals (tSsa1, tAdh1, tTdh1) were tested in place of the SV40 pA signal. Reporter expression remained similar across all pA sequences (Figure S3A), indicating their interchangeability.
• Promoter Tuning (Promoter Strength): A set of mutant T7 RNAP promoters with known decreasing strengths (characterized in vitro and in E. coli) were used to demonstrate tunable transcriptional response.
  • Cargo Independence: ZsGreen was swapped with other fluorescent proteins (BFP and mScarlet-I). While expression levels varied slightly, the high expression and tunability were maintained for all cargos (Figures 2A, 2B, S4A), indicating cargo independence.
  • Predictable Graded Expression: Three promoter variants provided graded levels of attenuated transcriptional activity. The rank order of expression levels was conserved, demonstrating predictable control over gene expression (Figures 2C, 2D, S4B).
• Multiplexed Gene Expression: A genetic circuit was designed with three separate fluorescent protein cargos (ZsGreen, mScarlet-I, BFP), each controlled by its own T7 promoter (Figure 2E). The T7 promoter for ZsGreen was systematically varied using the expression-level-modulating mutant versions, while the others remained constant. Normalized gene expression showed that only ZsGreen levels varied according to its promoter strength, while mScarlet-I and BFP remained constant (Figure 2F). This demonstrated independent control of multiple genes from a single, orthogonal master regulator.
• Orthogonal NPT7 Variants: Five orthogonal variants of $v443$ were created by grafting specific DNA-binding domain mutations (previously characterized for T7 RNAP to recognize distinct promoters). Each variant was tested against its cognate promoter driving a target gene (Figure 3A). The results showed successful switching of promoter specificity, but protein production levels were lower than the wild-type NPT7, suggesting a need for further optimization of these specific orthogonal fusions (Figure 3B).

4.2.5. Validation in Mammalian Cells

• Cross-Kingdom Portability: $v433$ and $v443$ were characterized for cytoplasmic activity in HEK293T mammalian cells.
  • The variants were cloned without their NLS tags (for cytoplasmic activity) and controlled by the CAG promoter and the beta-globulin polyadenylation signal.
  • Reporter expression was measured using two different reporter contexts with varying 5' and 3' UTR sequences (previously used in mRNA vaccine candidates).
  • The evolved variants showed a 3- to 4-fold higher reporter expression compared to the wild-type NPT7 in both contexts (Figures 4A-4C). This demonstrated that the yeast-evolved variants could function with enhanced activity in a mammalian system, highlighting their broad applicability.

5. Experimental Setup

5.1. Datasets

The study primarily utilizes two experimental model systems:

• Saccharomyces cerevisiae BY4741: This is a well-characterized strain of budding yeast, a common eukaryotic model organism.

  • Source: ATCC (American Type Culture Collection), identifier 201388.
  • Characteristics: It's a derivative of S288C and has the genotype $MATa his3Δ1 leu2Δ met15Δ0 ura3Δ0$. This background is common for genetic manipulations and reporter assays.
  • Domain: Fungal eukaryotic system.
  • Usage: Used as the primary host for developing and evolving the NPT7 fusion enzyme due to its ease of genetic manipulation, fast growth, and suitability for directed evolution platforms (e.g., FACS sorting of fluorescent populations). All initial evolution, tunability, and orthogonality characterizations were performed in yeast.
  • Data Sample: The paper uses ZsGreen, BFP, and mScarlet-I fluorescent proteins as reporters. A data sample would involve measuring the fluorescence intensity from yeast cells expressing these proteins. For example, higher ZsGreen fluorescence directly correlates with higher NPT7 activity.

• HEK293T Cells: A human embryonic kidney cell line.

  • Source: ATCC, identifier CRL-3216.

  • Characteristics: These cells are widely used in molecular biology for their ease of transfection and high protein expression levels. They represent a mammalian eukaryotic system.

  • Domain: Mammalian eukaryotic system.

  • Usage: Used to validate the cross-kingdom portability and enhanced activity of the yeast-evolved NPT7 variants in a clinically relevant context.

  • Data Sample: Similar to yeast, ZsGreen and mScarlet-I fluorescent proteins were used as reporters in HEK293T cells, with fluorescence intensity serving as the readout for NPT7 activity.

    These datasets were chosen because S. cerevisiae provides a robust and tractable platform for directed evolution and initial characterization of eukaryotic genetic circuits, while HEK293T cells offer a relevant mammalian model to test the broader applicability of the engineered system. This combination effectively validates the method's performance across different eukaryotic domains.

5.2. Evaluation Metrics

The primary evaluation metric used throughout the paper is reporter gene expression, which is quantified by measuring the fluorescence intensity of various fluorescent proteins.

1. Conceptual Definition: Reporter gene expression, in this context, refers to the level at which a specific gene (the reporter, e.g., ZsGreen, BFP, mScarlet-I) is transcribed and translated into a functional protein, whose presence can be easily detected and quantified. The design goal is that the amount of fluorescent protein produced is directly proportional to the activity of the NPT7 fusion enzyme. Higher fluorescence indicates higher transcriptional and capping activity of the engineered enzyme, leading to more functional mRNA and subsequent protein synthesis.

2. Mathematical Formula: The paper does not provide a specific mathematical formula for calculating "reporter expression" or "activity." Instead, it relies on the direct measurement of fluorescence intensity using flow cytometry (spectral analyzer SA3800 from Sony) and reports results as fold induction relative to a control (e.g., wild-type or background). The concept of fold induction is generally calculated as: $ \text{Fold Induction} = \frac{\text{Mean Fluorescence Intensity (Experimental Sample)}}{\text{Mean Fluorescence Intensity (Control Sample)}} $

3. Symbol Explanation:

   • Mean Fluorescence Intensity (Experimental Sample): The average fluorescence signal measured from cells expressing the reporter gene under the influence of the NPT7 variant being tested.

   • Mean Fluorescence Intensity (Control Sample): The average fluorescence signal measured from a control group of cells. This control could be cells expressing a wild-type NPT7, cells with a non-functional NPT7 (e.g., capping-null mutant), or cells without any NPT7 expression (background).

     The paper also uses statistical significance tests ($p < 0.05$ is *, $p < 0.01$ is **, and $p < 0.001$ is ***) to compare the means of different experimental groups, indicating the reliability of the observed differences.

5.3. Baselines

The paper compares its engineered NPT7 variants against several baselines to demonstrate improvement and functionality:

• Wild-type (WT) NPT7 Fusion Enzyme: This is the primary baseline. The engineered NPT7 without any directed evolution is used to show the significant enhancement in activity achieved by the evolved variants ($v433$ and $v443$). For instance, $v433$ and $v443$ showed nearly two orders of magnitude higher activity compared to the WT NPT7 (Figure 1C).

• Capping-Null NPT7 Variants: Mutated versions of both WT and evolved NPT7 (specifically, K294A in NP868R) that lack capping activity serve as crucial negative controls. These baselines confirm that the observed protein expression is indeed dependent on the capping function of the fusion enzyme (Figure S1B).

• Host-Regulated Promoters (Endogenous Gal promoter): In yeast, the activity of $v443$ is compared against the Gal promoter, a strong, endogenous, galactose-inducible promoter. This comparison provides a benchmark for how well the orthogonal NPT7 system performs relative to native eukaryotic expression systems (Figure 1H).

• Unlinked T7 RNAP and Capping Enzyme Co-expression: Although not directly re-evaluated in this paper, the introduction section explicitly states that previous studies showed enhanced protein expression with fusion enzymes compared to unlinked enzymes (Ref. 28). The current work builds on this prior finding by further optimizing the fusion.

• Previous T7 RNAP systems in eukaryotes: Generally, the paper positions its work as an advancement over prior attempts to adapt T7 RNAP for eukaryotic protein expression, which were limited by the lack of 5' capping or low expression levels.

  These baselines are representative because they either represent the starting point of the current engineering effort (WT NPT7), demonstrate the mechanism of action (capping-null mutants), or provide a benchmark against which the performance of the orthogonal system can be meaningfully evaluated (host promoters).

6. Results & Analysis

6.1. Core Results Analysis

The study's results demonstrate the successful engineering and significant enhancement of an orthogonal transcription engine for eukaryotes, highlighting its programmability, tunability, and portability.

Initial NPT7 Activity in Yeast and Capping Dependency: The wild-type NPT7 fusion enzyme, when introduced into Saccharomyces cerevisiae, showed only a modest ~2-fold increase in reporter protein production upon galactose induction (Figures 1B and 1C). This confirmed that while the fusion enzyme could function, its activity was low. Crucially, a capping-null mutant (K294A in NP868R) completely abolished this activity (Figure S1B), unequivocally demonstrating that the observed protein expression was dependent on the capping function of the NP868R domain within the fusion. This initial low activity served as the motivation for directed evolution.

Dramatic Enhancement via Directed Evolution: Directed evolution proved highly effective. After several rounds of selection, two variants, $v433$ and $v443$, were isolated. These variants exhibited an astounding nearly two orders of magnitude (approximately 100-fold) increase in reporter expression compared to the wild-type NPT7 (Figure 1C). Genotyping revealed 15 mutations in $v433$ and 10 in $v443$, with convergence at five residues, suggesting these specific mutations contribute significantly to the enhanced activity (Figure 1D; Table S1). Although capping-null versions of the evolved variants still showed about 10-fold higher reporter activity than the wild-type inactivated NPT7, this indicates a low level of background capping activity or non-capping related enhancements, but the vast majority of the evolved activity still relied on the capping function (Figure S1B). The evolved NPT7 did not significantly activate genomic targets, likely due to repressive chromatin conditions (Figure S1C), leading the authors to use high-copy reporter plasmids for further characterization.

Tunable Control and Comparability to Host Promoters in Yeast: The evolved variants demonstrated excellent tunable control. By varying galactose concentrations to modulate the expression level of the NPT7 fusion enzyme, corresponding graded levels of reporter expression were achieved (Figure 1F). This highlights the system's ability to predictably control gene output based on enzyme input. Furthermore, the expression levels achieved by the highly active $v443$ variant were comparable to those driven by a strong endogenous yeast Gal promoter (Figure 1H), showcasing that the orthogonal system can match the efficiency of native host machinery. The nuclear activity of the enzymes was also confirmed by the significant reduction in reporter expression when the NLS was removed (Figure S2B).

Predictable Tuning with T7 Promoters and Cargo Independence: The system also allowed for predictable tuning of expression by using different T7 RNAP promoter variants known to have varying strengths. The rank order of promoter variants' activity was conserved across different fluorescent protein cargos (ZsGreen, BFP, mScarlet-I) (Figures 2A, 2B, 2C, 2D, S4A, S4B). This indicated that the enhanced activity of the fusion enzyme is not specific to a particular cargo and that the intrinsic properties of the T7 promoters are maintained, providing a robust and predictable control mechanism orthogonal to host machinery. The interchangeability of different endogenous yeast polyadenylation signals also simplified circuit design (Figure S3A).

Multiplexed and Orthogonal Gene Expression: A key strength demonstrated was the ability to achieve multiplexed gene expression. A circuit expressing three different fluorescent proteins, each under its own T7 promoter, allowed for independent control. When the promoter strength for ZsGreen was varied, only its expression changed, while mScarlet-I and BFP levels remained constant (Figure 2E and 2F). This proves that a single master regulator enzyme can provide independent control over multiple target genes. The paper also explored truly orthogonal T7 RNAP variants by grafting specific DNA-binding domains to $v443$. These variants successfully switched promoter specificity, although with lower overall protein production than the wild-type T7 RNAP, suggesting further optimization is needed for these specific orthogonal fusions (Figures 3A and 3B).

Enhanced Performance in Mammalian Cells: Despite being evolved in yeast, the engineered variants $v433$ and $v443$ showed enhanced performance in HEK293T mammalian cells. They exhibited a 3- to 4-fold higher reporter expression compared to the wild-type NPT7 in two different reporter contexts (Figures 4A-4C). This demonstrates the crucial cross-kingdom portability of the evolved enzyme, making it broadly applicable beyond the model organism in which it was optimized.

In summary, the results strongly validate the effectiveness of the proposed method. Directed evolution dramatically improved the NPT7 fusion enzyme, overcoming a major bottleneck for T7 RNAP systems in eukaryotes. The system offers unparalleled orthogonality, tunability, and programmability, distinguishing it from previous attempts and providing a powerful tool for synthetic biology applications. While the activity of orthogonal variants was lower and nuclear activity on genomic targets was limited, the overall advancements are substantial.

6.2. Data Presentation (Tables)

The following are the results from KEY RESOURCES TABLE of the original paper:

6.2.1. Analysis of Figures

The figures effectively visualize the experimental results, providing strong evidence for the paper's claims.

The following figure (Figure 1 from the original paper) illustrates the directed evolution process and the initial characterization of NPT7 variants in yeast.

该图像是图表，展示了在酿酒酵母中对NPT7变体的定向进化和特征化。图中包括不同变体（如v433和v443）在诱导下的表达活性，数据以平均值ext{±}标准差表示，反映了对合成生物学应用的可调控控制。

As seen in Figure 1, panel A, the system involves an NPT7 fusion enzyme with an NLS integrated into the yeast genome, and a ZsGreen reporter plasmid under a T7 promoter. Panels B and C clearly show the improvement from directed evolution, with $v433$ and $v443$ exhibiting significantly higher fluorescence (protein expression) compared to the WT NPT7. Panel D lists the specific mutations found in the evolved variants, supporting the directed evolution process. Panels E and F demonstrate the tunability of the system, where varying galactose concentrations (and thus NPT7 expression) lead to graded reporter expression. Panel H compares the expression from $v443$ to endogenous Gal promoters, showing comparable activity.

The following figure (Figure 2 from the original paper) demonstrates the tunable control of gene expression using NPT7 with mutant T7 RNAP promoters and the system's cargo independence.

该图像是示意图，展示了不同基因在酵母和哺乳动物细胞中的折叠诱导。图A展示了转录调控机制，图B和D显示了v443和v433的基因表达折叠诱导的量化数据，而图F则比较了不同基因在BFP表达下的相对表达水平。

Figure 2 illustrates that the evolved NPT7 variants ($v433$ and $v443$) can predictably tune gene expression. Panels A and B show that different fluorescent protein cargos (BFP, mScarlet-I, ZsGreen) can be expressed, and the high expression level is maintained. Panels C and D demonstrate that a range of mutant T7 promoters with varying strengths can reliably provide graded levels of gene expression, maintaining their rank order of activity. This confirms the predictable and orthogonal control. Panels E and F further highlight multiplexed expression, where ZsGreen expression can be independently varied using different promoter strengths, while mScarlet-I and BFP expression remains constant, all driven by a single master NPT7 enzyme.

The following figure (Figure 3 from the original paper) characterizes the mutually orthogonal NPT7 variants.

该图像是图表，展示了互相正交的NPT7变体的表征。在图A中，显示了不同的表达载体及其对应的荧光蛋白ZsGreen的表达模式。图B展示了各变体的折叠诱导度，表格中包含多种变体的折叠诱导量，显示不同表达系统的有效性。

Figure 3, panel A, outlines the design of five orthogonal $v443$ variants created by grafting specific DNA-binding domains. Panel B presents the results, showing that these variants successfully recognize their cognate promoters, enabling specific "switching" of promoter specificity. While the protein production levels were noted to be lower than the WT NPT7, this experiment validates the concept of developing multiple, independent orthogonal channels within the capping-T7 system.

The following figure (Figure 4 from the original paper) shows the characterization of the evolved NPT7 variants in HEK293T cells.

该图像是图表，展示了在HEK293T细胞中对进化型NPT7变体的表征。图中包括两个荧光蛋白ZsGreen和mScarlet-I的表达结果，以及不同的5'和3' UTR序列对其表达的影响。通过生物重复实验，结果以均值±标准差（SD）呈现，归纳了不同变体在荧光强度上的表现。

Figure 4 presents evidence of the cross-kingdom portability and enhanced activity of the evolved NPT7 variants in mammalian cells. Panels A, B, and C show that $v433$ and $v443$ (without NLS) exhibit a 3- to 4-fold higher reporter expression (for both ZsGreen and mScarlet-I) compared to the WT NPT7 in HEK293T cells, across different 5' and 3' UTR contexts. This is a crucial finding, demonstrating that the yeast-evolved improvements translate to mammalian systems.

6.3. Ablation Studies / Parameter Analysis

The paper includes several elements that function as ablation studies or parameter analyses, verifying the effectiveness of specific components and parameters:

• Capping-Null Mutants (Ablation of Capping Function): The introduction of the K294A point mutation in the NP868R domain, creating a "capping dead" version of NPT7, serves as a direct ablation study.

  • Results: The wild-type capping-null NPT7 showed significantly reduced reporter expression, confirming the critical role of the capping activity (Figure S1B). Similarly, capping-null versions of the evolved $v433$ and $v443$ variants also showed drastically reduced activity compared to their functional counterparts, validating that the enhanced activity was still primarily due to improved capping function.
  • Effectiveness: This confirms that the primary mechanism by which NPT7 enables protein expression is through 5' capping.

• Nuclear Localization Signal (NLS) Removal (Ablation of Nuclear Targeting): The authors tested NPT7 variants without the NLS tag.

  • Results: Removal of the NLS led to a significant reduction in reporter expression (Figure S2B).
  • Effectiveness: This confirms that the enzyme's activity is indeed nuclear in yeast, where the target DNA plasmid is located, and that the NLS is critical for proper localization and function.

• Varying Galactose Concentration (Parameter Analysis for Enzyme Level): The use of a $ΔGal2$ yeast strain and varying levels of galactose in the culture medium allowed for precise control over the expression levels of the NPT7 fusion enzyme.

  • Results: Graded levels of reporter expression were obtained by varying galactose concentrations (Figure 1F).
  • Effectiveness: This demonstrates that the system is tunable and that the final gene expression can be regulated by controlling the amount of the fusion enzyme available.

• Mutant T7 RNAP Promoters (Parameter Analysis for Promoter Strength): The characterization of a set of T7 promoter variants with decreasing strengths.

  • Results: These promoters consistently provided graded levels of transcriptional activity, leading to proportional protein expression. The rank order of promoter variants was conserved across different circuit designs (Figures 2C, 2D, S4B).
  • Effectiveness: This shows that the intrinsic tunability of T7 promoters is maintained within the NPT7 system, allowing for predictable control over gene expression orthogonal to host machinery.

• Different Fluorescent Protein Cargos (Parameter Analysis for Cargo Specificity): The swapping of ZsGreen with BFP and mScarlet-I fluorescent proteins.

  • Results: While expression levels varied slightly, the high expression and tunability were observed for all cargos (Figures 2A, 2B, S4A).
  • Effectiveness: This indicates that the increased activity of the fusion enzyme variants is not cargo-specific, making the system broadly applicable for expressing various proteins.

• Different Yeast Polyadenylation Signals (Parameter Analysis for 3'UTR Elements): The systematic swapping of the SV40 pA signal with three different yeast signals (tSsa1, tAdh1, tTdh1).

  • Results: Reporter expression remained similar across all tested pA sequences (Figure S3A).

  • Effectiveness: This suggests that these sequences are largely interchangeable within the system, simplifying circuit design and increasing modularity.

    These analyses collectively demonstrate the robustness, predictability, and modularity of the engineered capping-T7 system, confirming that its key components and parameters function as intended.

7. Conclusion & Reflections

7.1. Conclusion Summary

This study successfully developed and significantly enhanced an orthogonal transcription engine for programmable gene expression in eukaryotes. By fusing T7 RNA polymerase with the NP868R viral capping enzyme to create NPT7, the researchers addressed the critical challenge of 5' capping for T7-derived transcripts in eukaryotic cells. Through an iterative directed evolution approach in Saccharomyces cerevisiae, they isolated variants ($v433$ and $v443$) that demonstrated an impressive ~100-fold increase in activity compared to the wild-type fusion enzyme. This evolved system provides highly programmable and tunable control over gene expression, maintaining orthogonality to host factors, and proving effective in both yeast and mammalian cells. The findings establish capping-T7 as a robust, portable, and scalable tool, significantly advancing synthetic biology applications in diverse eukaryotic hosts.

7.2. Limitations & Future Work

The authors candidly acknowledge several limitations and propose future research directions:

• Disparity in Activity Enhancement Across Hosts: While the evolved NPT7 variants showed a nearly 100-fold improvement in yeast, the enhancement in mammalian HEK293T cells was more modest (3-4 fold).

  • Limitation: This suggests the variants are not fully optimized for mammalian systems, potentially due to host-specific factors like nuclear import, post-translational modifications, or interactions with endogenous proteins.
  • Future Work: Further optimization campaigns specifically in mammalian cellular environments (and other eukaryotic hosts like plant cells or primary mammalian cells) are warranted to improve activity and assess generalizability.

• Lack of Individual Mutation Analysis: The evolved variants ($v433$ and $v443$) contain multiple amino acid substitutions (15 and 10, respectively).

  • Limitation: The specific contribution of individual mutations to the enhanced activity remains undetermined.
  • Future Work: Detailed structure-function analyses and mutational studies are needed to understand the underlying mechanisms of increased activity and to develop more general design rules for engineering such fusion enzymes in other chassis organisms.

• Untested Long-term Stability and Epigenetic Silencing: The study did not investigate the long-term stability of gene expression mediated by the evolved NPT7 variants.

  • Limitation: Eukaryotic cells can epigenetically silence exogenous genetic elements over time, which could diminish the system's effectiveness.
  • Future Work: Studies examining expression persistence and potential epigenetic modifications are essential to understand the long-term efficacy and stability of the orthogonal transcription system and to inform strategies to mitigate silencing.

• Limited Genomic Integration Activity: The evolved NPT7 did not show significant activity on genomic targets in yeast.

  • Limitation: This suggests that repressive chromatin conditions might hinder its function on integrated genes, potentially limiting its application for stable genomic expression.
  • Future Work: Further evolution could aim to improve function under repressive chromatin conditions, or alternative strategies like expressing the fusion enzyme in the cytosol using systems like OrthoRep could be explored.

7.3. Personal Insights & Critique

This paper presents a highly impactful advancement in eukaryotic synthetic biology. The integration of a viral capping enzyme with T7 RNAP is an elegant solution to a long-standing problem, but the real power comes from the application of directed evolution. This highlights a crucial lesson in biomolecular engineering: initial rational design might provide a functional scaffold, but iterative, high-throughput evolutionary optimization is often necessary to achieve truly high-performance biological tools.

Inspirations and Applications:

• Refactoring Metabolic Pathways: The ability to precisely tune and orthogonally control gene expression in yeast has significant implications for metabolic engineering. It could enable more stable and fine-grained flux tuning for the production of high-value compounds, mitigating issues of host regulation and epigenetic silencing.
• mRNA Therapeutics and Vaccines: The method offers a simplified manufacturing process for 5' capped mRNA, which is foundational to RNA therapeutics (like mRNA vaccines). Currently, T7 RNAP is used, but a separate capping step is required. A single-step capping-T7 enzyme could streamline production, reducing cost and complexity. The authors' note about further optimization for reduced dsRNA byproducts (as seen in recent work on T7 RNAP variants for mRNA synthesis, Refs. 49, 50) is also highly relevant here.
• Complex Genetic Circuitry: For fields aiming to build complex, multi-gene circuits in eukaryotes (e.g., for biosensors, cell programming, or developmental biology studies), this capping-T7 system provides a level of orthogonality and predictability previously unattainable, allowing for true bottom-up construction independent of confusing host interactions.

Potential Issues or Areas for Improvement:

• Mechanism of Evolution: While the paper identifies mutations, a deeper dive into how these mutations enhance activity (e.g., improved protein stability, increased catalytic rate, better interaction between T7 RNAP and capping enzyme domain, or altered substrate binding) would be highly valuable. This could lead to more rational design principles.

• Orthogonal Variant Performance: The orthogonal T7 RNAP variants showed lower protein production. Optimizing these variants through further directed evolution is a clear next step to fully realize multiplexed control.

• Long-Term Stability and Immunogenicity in Mammalian Systems: For therapeutic applications, the long-term stability of expression in mammalian cells and potential immunogenicity of a viral-derived fusion enzyme would be critical considerations, requiring extensive in vivo studies.

• Generalizability to Diverse Eukaryotic Hosts: While the paper shows portability to HEK293T cells, the performance in other important eukaryotic systems (e.g., primary cells, stem cells, plant cells, or different yeast species) remains to be thoroughly investigated. The 3-4 fold improvement in mammalian cells, while positive, is not as dramatic as the 100-fold in yeast, suggesting cell-specific context matters.

  Overall, this paper represents a significant technological leap forward, enabling more precise and robust engineering of eukaryotic gene expression. The combination of clever enzyme fusion with powerful directed evolution is a potent strategy for overcoming biological engineering challenges.