Engineering Success: the PHAGEVO Directed Evolution System

The PHAGEVO solution is an improved directed evolution technology that combines the strength of phage-assisted continuous evolution (PACE) and Evolution.T7, a continuous selection system with auto-amplification of the best fitted variants for the former [Liu et al., 2011] and a targeted in vivo mutagenesis system focusing evolution on a desired gene for the latter (https://2021.igem.org/Team:Evry_Paris-Saclay).

Based on the latest improvements of the PACE technology, notably the non-continuous version of it (PANCE) and the use of gene VI as a selection marker, we first focused on building the selection phages strain incorporating the Evolution.T7 cassette, a synthetic Escherichia coli chassis adapted for PHAGEVO, the accessory and mutagenesis plasmids required for a versatile PHAGEVO platform that could be used for in vivo directed evolution of a protein of interest.

For the proof of concept of our system, we all agreed on searching for a protein that could have a positive impact on sustainable development. The successful candidate was XylS, a transcription factor that was recently engineered toward detection of phthalic acid (PA) and the plastic degradation product terephthalic acid (TPA) [Li et al., 2022]. Our goal was to evolve two XylS mutants already reported to have the ability to detect of plastic degradation products (PA and TPA): K38R-L224Q and W88C-L224Q, in order to get better variants with optimized PA or TPA biosensing properties.

This page describes our strategic choices, our designs and our experimental results. It is organized in several chapters that explain:

  1. The background of the project, summarizing our knowledge on the PACE technology and on the M13 bacteriophage it is based on and on the Evolution.T7, both with their advantages and disadvantages
  2. The design of the PHAGEVO system and its experimental implementation

PHAGEVO Project Background

I.1 M13 Bacteriophage

Escherichia phage (E. coli) M13 bacteriophage is the most studied member of the Inovirus genus, part of the Inoviridae family, whose name is derived from the Greek word "ina," meaning fiber or filament (Knezevic et al., 2021). Inoviruses infect Gram-negative bacteria like Escherichia coli and have a characteristic elongated filamentous capsid that packages the genome, a circular single-stranded DNA (ssDNA) of positive sense, ranging from 5.5 to 10.6 kilobases (kb) in length (Figure 1).

The M13 genome (6.407 kb, GenBank V00604, Figure 2) encodes 11 different proteins, numbered from p1 to p11, with a size between 3.3 kDa and 50 kDa.

Their interplay during the bacteriophage life cycle hijacks cellular processes to produce new virus particles, as schematised in Figure 3.

M13 infection begins with the recognition and adhesion of the phage to the bacterial surface via the phage pIII protein and the bacterial F pilus (a transmembrane protein) and the TolQRA co-receptor located in the inner membrane of the Gram-negative bacterial host. This limits the M13 infection only to E. coli cells carrying a wild-type F plasmid or a modified F’ one, a feature we took into account in the choice of the E. coli host strain in our PHAGEVO project (see details below). Following this interaction, the M13 ssDNA is translocated into the cytoplasm where it remains in a circular, extra-chromosomal form, similar to a plasmid, without integrating into the host genome. Phage DNA is then transcribed into mRNA by RNA polymerase III, and replicated following a rolling-circle mechanism. The newly synthesized ssDNA is stabilized through interaction with protein pV, which is later replaced by pVIII, the major coat protein, during virion assembly and extrusion. In this process, the mature virions are released into the environment starting with the pVII and pIX end, followed by multiple copies of pVIII, while pIII and pVI are added at the final stage.

M13 establishes a chronic productive infection, where virions are continuously released from infected cells via extrusion, while the cells remain viable and intact, although typically exhibiting a slower growth rate.

In the absence of pIII protein, the phage loses its ability to recognize and infect host bacteria, while in the absence of pVI phage particles remain associated with the host membrane. Our PHAGEVO technology leverages this feature by making pIII and/or pVI selection markers for phage production depending on the activity of XylS, our protein of interest.


Figure 1. Schematic representation of the M13 bacteriophage. pIII, pVI, pVII, pVIII and pIX are coat proteins that form the virion and encapsulate the single stranded circular DNA (ssDNA). The virion consists of non-enveloped, flexible filaments, measuring between 0.6 and 2.5 µm in length and 6-7 nm in diameter. Its length varies depending on the size of the packaged DNA, which is an adjustable variable. The major coat protein, pVIII (CoaB), forms the outer layer of the filament, along with proteins pVII and pIX, while the inner coat is composed of pVI. The pIII protein (CoaA) plays a crucial role in adhesion, recognition, and the infection process. The 3D structure of the M13 virion is a helical complex, where pVIII forms an alpha-helix that makes up the viral coat. The DNA inside the virion is supercoiled and partially folded into a double-helix structure. It adopts both A-form and B-form conformations of double-stranded DNA (dsDNA), with the top segment in an unfolded, unwound state, the middle segment in A-form DNA, and the bottom segment in the more common B-form, which is typical of many protein-DNA complexes. (adapted from Knezevic P et al., 2021)

Figure 2. M13 bacteriophage genome map (GenBank: V00604)

Figure 3. Schematic representation of the infection and replication cycle of the M13 bacteriophage. Image adapted from Knezevic et al., 2021.

I.2 PACE PANCE PRANCE EVOLUTION SYSTEMS

I.2.1 P A C E

Phage-Assisted Continuous Evolution

PACE is a continuous directed-evolution system based on the M13 bacteriophage [Esvelt et al., 2011; Miller et al., 2020a].
As described above, infectivity of M13 phages is dependent on filamentous phage protein III encoded by gene III (pIII / gIII), that binds specifically to F-pilus for entry into the bacterial cytoplasm. In PACE, the phage genome is depleted from gIII which is replaced by the gene of the protein to evolve (protein of interest, POI) (Selection phage, SP Figure 4). gIII is expressed from another plasmid (accessory plasmid, AP on Figure 4). In order to evolve the POI toward the desired function, the expression level of gIII must be dependent on POI and linked to an increase in POI fitness. Therefore, the only phages to harbor gIII on their capsid (and therefore only infective phages) will be the ones carrying a genome (Selection phage, SP Figure 4) with an active POI. Deleterious mutations of the POI will be eliminated because phages carrying inactive POI will lack gIII and will be unable to infect bacteria and multiply. Over several cycles of infection, the medium is enriched in phages expressing a POI with beneficial mutations thanks to the natural selection process between the different phages. In addition, to increase mutation rates, host bacterial cells carry a mutagenesis plasmid (MP in Figure 4). The most potent one is MP6 which expresses dnaQ926, dam, seqA, emrR, ugi and PmCDA1 that insert mutations, block the DNA repair system and are involved in DNA methylation system [Badran and Liu, 2015].

Infectivity of M13 phages is also dependent on the protein VI encoded by gene VI (pVI / gVI) which was developed as an alternative selection marker [Brödel et al., 2016; Miller et al., 2020b] especially because low levels of pIII expression make the host cell resistant to infection [Boeke et al., 1982].

PACE operates in a continuous manner through repeated cycles of bacterial infection and POI’s activity-based bacteriophage production. For this, fresh host bacterial cells are continuously introduced into a reaction vessel called a lagoon, where phages infect the bacteria, replicate, and then go on to infect new bacterial cells. This continuous system is set up in a custom-built PACE apparatus, where the inflow of fresh cells into the lagoon and the outflow mechanism for removing bacteria are regulated at rates carefully adjusted to allow phage replication within the lagoon without washing away all the phages.



Figure 4. The principle of PACE and PANCE and PRANCE systems for continuous, non-continuous and robotics-assisted near-continuous evolution with SP for selection phage, AP for accessory plasmid, MP for mutagenesis plasmid and POI for protein of interest. The red dots correspond to mutations introduced into the genome and plasmids. Phages are periodically isolated and introduced in a medium with new host cells (adapted from Bradan et al., 2016 ; Brödel et al., 2018 ; Esvelt et al., 2011; Miller et al., 2020a).

I.2.2 P A N C E

Phage-Assisted Non-continuous Evolution

PANCE is a non-continuous version of the PACE system, operating on the same principles but requiring human intervention between each evolution cycle: instead of a continuous supply of host cells as in PACE, in PANCE a subculture is initiated in fresh media with fresh host cells with phages isolated from the previous culture [Miller et al., 2020]. This discontinuous process makes PANCE slower than PACE in terms of culture growth and mutant appearance. However, the manual selection of variants allows for better control over each evolution cycle, including the selection of phages and variants, despite a higher potential for bias. PANCE is easier to implement and more flexible for various lab environments, as it doesn’t require custom-made continuous flow bioreactors systems. It is more adaptable and enables multiple evolution experiments to be conducted in parallel in 96-well plates for instance.

I.2.3 P R A N C E

Phage and Robotics-Assisted Near-Continuous Evolution

PRANCE is a high-throughput system developed to automate the PACE process, traditionally performed in a bioreactor, by utilizing 96-well plates and robotic liquid-handling equipment [DeBenedictis et al., 2022]. This setup allows for hundreds of independent evolution experiments to be run simultaneously, and, by integrating real-time monitoring of biomolecular activity, automated feedback control mechanisms, high-throughput sequencing, PRANCE allows precise control over environmental factors such as media composition, OD600, timing, and chemical conditions, enabling experiments under varied conditions, including different temperatures, host strains, and chemical environments.

I.2.4 ADVANTAGES and DISADVANTAGES of PACE / PANCE / PRANCE EVOLUTION SYSTEMS

The elegance of the PACE system and its declinations is that it makes an auto-screening of the best variants that have an improved activity of the protein to evolve (POI). Indeed, as the expression levels of the essential gIII (or gVI) are regulated by the POI, the infective phage production levels are also tightly regulated by the POI. This results in an auto-amplification of the best variants, that will quickly outcompete the ineffective variants in the lagoon where evolution occurs. With time, the POI will continuously evolve toward the optimal sequence for the desired properties, and at the end of the experiment only a limited number of variants with increased fitness will be found in the lagoon.

As for all technologies, PACE has limitations. The major challenge when designing a PACE experiment is to be able to link the activity of the POI to the expression levels of gIII (or gVI). This is essential for PACE as it provides the selection pressure toward the best adapted POI. PACE, PANCE, and PRANCE have all been successfully employed to engineer a wide range of proteins. The easiest targets for evolution using PACE, and the first to be evolved with this technology, are proteins directly involved in gene regulation, such as T7 RNA polymerase [Liu et al., 2011], transcription factors [Brödel et al., 2016] or DNA-binding proteins like TALEN [Hubbard et al., 2015]. More recently, other types of proteins have been engineered with PACE. This includes biosensors for small molecules ([Jones et al., 2021], [Li et al., 2022]), enzymes like proteases [Dickinson et al., 2016] or involved in metabolic pathways like the methanol dehydrogenase Bm Mdh2 [Roth et al., 2019], the therapeutic target PD1 involved in cancer immune escape [Ye et al., 2020] or protein-protein interaction [Wang et al., 2018].

The latest improvements of the PACE technology includes PRANCE [DeBenedictis et al., 2022], which enables the high-throughput study of factors influencing evolution, and makes it possible to study how the conditions of evolution influence evolutionary trajectories of the gene of interest. PRANCE is particularly useful for protein or RNA engineering and enables the study of evolutionary dynamics in a controlled environment which facilitates the analysis of complex evolutionary processes.

In PACE, PANCE and PRANCE, mutagenesis is triggered by a mutagenesis plasmid expressing error-prone DNA polymerase, inhibit DNA repair pathways, increases DNA methylation or inhibit elimination of mutagenic nucleobase [Miller et al., 2020]. The most potent mutagenesis plasmid increases mutation rates 300 000 fold, reaching approximately 6.10⁻⁶ mutations per base pair per generation. However, these mutations are not restricted to the gene of interest and occur on the whole genome. Therefore, mutations can accumulate on the phage genome and lead to defectuous phages and loss of POI mutants. In PANCE, where bacteria are not continuously renewed, mutations on the genome can also lead to mortality or defects in the plasmids required for PACE.

To further increase mutation rates and fasten PACE, PANCE or PRANCE experiments, we decided in PHAGEVO to target evolution on the gene of interest only. Doing this may not only further increase mutation rates, but also reduce failure rate of the experiment and phage washout, which is a common issue arising during PACE, by preventing deleterious mutations on the phage or bacteria genome.

I.3 The EVOLUTION.T7 System

Evolution.T7 is a tool developed by the iGEM Evry Paris-Saclay 2021 team [https://2021.igem.org/Team:Evry_Paris-Saclay]. It is based on the orthogonal T7 RNA polymerase (T7RNAP) linked to a base deaminase (BD) either a cytosine or an adenosine deaminase (respectively CD and de notre ère), which allows for the rapid generation of genetic diversity in GOI in vivo in E. coli. When BD-T7RNAP fusion protein is expressed, the sequence flanked by the T7 promoter and the T7 terminator(s) gets mutated as the CD or de notre ère randomly deaminates the nucleotides mainly on the non template strand of the T7RNAP. Upon DNA replication, these deaminated bases lead to C→T or A→G transition mutations, depending on whether CD or de notre ère was used (Figures 5 and 6).

MutaT7 was the first BD-T7RNAP-based tool reported in the literature [Moore et al., 2018]. It was using only the rApo1 / rAPOBEC1 cytosine deaminase and thus had a limited mutational spectrum, but, compared to existing direct mutagenesis tools, this method showed improved on-target mutagenesis. This advantage was the key reason for its subsequent development and the emergence of more advanced versions like T7-DIVA [Álvarez et al., 2020], eMutaT7 [Park & Kim 2021], dT7-Muta [Ting & Ng 2023], T7-DualMuta [Wei et al., 2023], Optimized Muta-T7 [Mengiste et al., 2023], eMuta-T7transition [Seo et al., 2023]. The primary host organism for these studies has been E. coli, but the system proved its functionality also in eukaryotic cells like mammalian [Chen et al., 2020], yeast [Cravens et al., 2021, Huang et al., 2023] or plant cells [Butt et al., 2022]. This versatility is on account of the orthogonality of the T7RNAP which is highly specific to its promoter sequence (TAATACGACTCACTATA) and is able to operate in other organisms than E. coli cells [Imburgio et al., 2000].


Figure 5. The mutation mechanism of the Evolution.T7 system with a base deaminase (BD) fused to the T7 RNA polymerase (T7RNAP) (adapted from Moore et al., 2018).


Figure 6. Mutation Mechanisms through Deamination. (A-C) In DNA, deamination of a cytosine by a cytosine deaminase converts it to deoxyuridine which pairs with adenosine and leads to a C > T mutation, and, when the deamination occurs on the reverse strand, a G > A mutation occurs. (D-F) Deamination of an adenine by an adenine deaminase converts it to deoxyinosine which pairs with C and causes an A > G mutation, and, when the deamination occurs on the reverse strand, a T > C mutation occurs.

Compared to other T7RNAP-based tools mentioned above, to be able to introduce also T→C and G→A mutations, Evolution.T7 uses also a mutated T7RNAPCGG-R12-KIRV specific to an altered T7CGG promoter sequence which was placed in the reverse orientation downstream of the target region in order to compensate for the above mentioned bias of deaminations occurring mainly on the non template strand (Figure 7).
Evolution.T7 is versatile, allowing for the adjustment of mutation rates by using different CD or de notre ère combinations. The system comprises five CDs (AID, pmCDA1, rAPOBEC1, evoAPOBEC1-BE4max, evoCDA1-BE4max) and two ADs (TadA* and ABE8.20m), each fused to T7RNAP and T7RNAPCGG-R12-KIRV. These components are carried on low-copy plasmids (pSEVA221 and pSEVA471) to reduce replication burden, minimize T7RNAP toxicity, and limit off-target mutations in the E. coli genome. Expression is inducible by anhydrotetracycline for BD-T7RNAP and by L-arabinose for BD-T7RNAPCGG-R12-KIRV, enabling either sequential or concurrent mutagenesis on both DNA strands.


Figure 7. Schematic of the general organization of the Evolution.T7 system. The GOI is flanked upstream and downstream by the PT7 (sense) and PT7CGG (antisense) promoters, respectively, and by four T7 terminators (BBaB0015, Sba000587, T7wt, Sba_000451).

PHAGEVO = PANCE x Evolution.T7

PACE is a powerful directed evolution tool notably due to its auto selection process of the best variants of the gene of interest (GOI) and their subsequent amplification. However, it finds its limit through the fact that mutations can occur everywhere in the plasmids, bacterial chromosomes and on the phage genome and lead to defects in essential genes, leading to the loss of mutants.
To contain and focus these mutations only into the GOI, and therefore considerably increase the recovery of new variants and mutagenesis rates, we combined PACE to a targeted evolution system. This targeted evolution system is Evolution.T7 [https://2021.igem.org/Team:Evry_Paris-Saclay] which was developed by the 2021 iGEM Evry Paris-Saclay team. Evolution.T7 is based on the orthogonal T7 RNA polymerase (T7RNAP) linked to a base deaminase (BD) either a cytosine or an adenosine deaminase (respectively CD and de notre ère). Owing to two different T7 promoters flanking the GOI and working in opposite direction, Evolution.T7 allows for the rapid generation of genetic diversity in GOI in vivo in E. coli with a limited mutation bias.

Our team took advantage of the best aspects of both technologies to design the PHAGEVO system. As summarized in Figure 8, several modifications were implemented in PHAGEVO compared to the PA(N)CE system. On the selection phage (SP), the GOI was flanked by the two T7 promoters (PT7 and PT7CGG) and the terminators from Evolution.T7. The mutagenesis plasmid was completely remodeled to accommodate the specificity of the Evolution.T7 targeted mutagenesis and express the T7 RNA polymerase linked to base deaminase. Only the accessory plasmid, which expresses pIII or pVI phage proteins under the control of the gene to evolve, remains unchanged in PHAGEVO compared to PA(N)CE.


Figure 8. PHAGEVO evolution system (adapted from Bradan et al., 2016; Brödel et al., 2018; Esvelt et al., 2011; Miller et al., 2020a)

In the development of our project we followed the Design, Build, Test, Learn (DBTL) cycle on each aspect. In the following parts of this page, we present how we gradually implemented it at each step:

II.1 Construction of E. coli S2060 ∆ung ∆nfi, a strain adapted for the PHAGEVO system

A question that arose early on in the development of our project was the choice of the E. coli host strain. Which one would be better for PHAGEVO between the S2060 developed for PACE and MG1655* Δflu ΔpyrF Δung Δnfi used in Evolution.T7 ?

II.1.1 S2060 versus MG1655* Δflu ΔpyrF Δung Δnfi

S2060 is a stain developed by Hubbard et al., 2015 as an improvement of earlier E. coli strains specifically engineered for PACE, S1030 [Carlson et al., 2014] and S109 [Esvelt et al., 2011], all of which were derived from the widely used and well-known DH10B [Durfee et al., 2008]. The modifications gradually introduced in DH10B are highlighted in Table 1. They are located both at chromosome level, as well as on the F’ plasmid. The main reasons for these modifications are reducing the biofilm formation which represents an important issue in continuous cultures, enabling a robust induction of the mutagenesis or the evolution of specific genes:

In addition, DH10B was modified to become an F' strain, an essential condition for M13 phage propagation as described above. The original F’ plasmid was acquired by conjugation from the E. coli ER2738 strain and had the genotype F´proA+B+ lacIq Δ(lacZ)M15 zzf Tn10(TetR). It was further modified to keep or acquire new features:

Table 1. The genotype of E. coli S2060 strain and of its ancestors. Modifications highlighted in blue are those introduced early on in the DH10B strain to construct the S109 strain, those in purple were introduced in the S109 to construct S1030, while those in red were added in S2060.

STRAIN GENOTYPE REFERENCES
DH10B F- endA1 recA1 galE15 galK16 nupG rpsL ΔlacX74 ϕ80dlacZΔM15 araD139 Δ(ara,leu)7697 mcrA Δ(mrr-hsdRMS-mcrBC) λ– Durfee et al., 2008
S109 F’ proA+B+ Δ(lacIZY) zzf::Tn10(TetR) / endA1 recA1 galE15 galK16 nupG rpsL ΔlacIZYA araD139 Δ(ara,leu)7697 mcrA Δ(mrr-hsdRMS-mcrBC) proBA::pir116 λ– Esvelt et al., 2011
S1030 F’ proA+B+ Δ(lacIZY) zzf::Tn10(TetR) lacIQ1 PN25-tetR luxCDE / endA1 recA1 galE15 galK16 nupG rpsL ΔlacIZYA araD139 Δ(ara,leu)7697 mcrA Δ(mrr-hsdRMS-mcrBC) proBA::pir116 araE201 ΔrpoZ Δflu ΔcsgABCDEFG ΔpgaC λ– Carlson et al., 2014
S2060 F’ proA+B+ Δ(lacIZY) zzf::Tn10(TetR) lacIQ1 PN25-tetR luxCDE Ppsp(AR2) lacZ luxR Plux groESL / endA1 recA1 galE15 galK16 nupG rpsL ΔlacIZYA araD139 Δ(ara,leu)7697 mcrA Δ(mrr-hsdRMS-mcrBC) proBA::pir116 araE201 ΔrpoZ Δflu ΔcsgABCDEFG ΔpgaC λ– Hubbard et al., 2015

E. coli MG1655* Δflu ΔpyrF Δung Δnfi is a stain developed by Álvarez et al., 2020 to increase the mutation rate in BD-T7RNAP-based mutagenesis tool. This strain was derived from the wild-type E. coli MG1655 in which several modifications have been introduced:

Heaving all these considerations in mind, we decided to take the best from the two strains and construct the S2060 Δung Δnfi. Indeed, the numerous modifications implemented in the S2060 strain make the monitoring of phage infection and thus of the evolution experiment easier to follow. In addition, its genetic background (DH10B) is that of a ‘cloning’ E. coli strain commonly used in laboratories as a competent cell for its high large plasmid DNA transformation efficiency and the construction of large DNA libraries. DH10 was constructed from MC1061 by introducing the recombinase-deficient mutant recA1, which reduces recombination with host DNA, and the endonuclease-deficient mutant endA1 mutation, which improves plasmid DNA quality. The strain also carries mutations in mcrA, mcrB, blocking the restriction of methylated cytosine DNA and deoR, a mutant allowing for the efficient propagation of large plasmids thanks the constitutive expression of genes for deoxyribose synthesis enabling deoxyribose synthesis [Durfee et al., 2008].

In Evolution.T7 and related T7RNAP-based tools, Δung and Δnfi are important modifications for enhancing the mutation rate. Without these genes, the DNA repair mechanisms are impaired both in case of cytosine and adenine deamination, and thus, the mutations generated are maintained.
Reducing the capacity of S2060 to repair mutation may be a problem in PACE / PANCE / PRANCE as mutations with the MP6 plasmids are introduced all over the genome. However, in PHAGEVO, mutations are targeted to the GOI, thus limiting the side-effects.

II.1.2 Construction of E. coli S2060 Δung Δnfi through CRISPR/Cas9 and λ Red mediated recombineering

II.1.2.1 Design

CRISPR/Cas9 is a powerful genome engineering technology, which makes it our first choice for constructing the S2060 Δung Δnfi.
CRISPR/Cas9 relies on the activity of the Cas9 endonuclease, which, when paired with a guide RNA (gRNA), allows for the specific targeting of a DNA sequence and induces a double-strand break. In the presence of a repair template (RT), an allelic substitution can occur through homologous recombination at the DNA break site, facilitated by the expression of a DNA repair system.
DNA cleavage occurs 3 bp upstream of the Protospacer Adjacent Motifs (PAM) (5’-NGG-3’) upon a stable interaction between the DNA and gRNA upstream of the PAM. Thus gRNAs with a 20 bp sequence complementary to the DNA upstream of a PAM can induce RNA-guided double-strand break in the target DNA, like a chromosome.

Using CHOPCHOP (version 3) web tool [Labun et al., 2019; Labun et al., 2016; Montague et al., 2014] we designed CRISPR/Cas9 target sites for both ung and nfi genes. For ung, we selected as target sequence AATCAGTGGCTGGAACAACGTGG, as the best ranked among 94 proposed for knocking-out this gene in E. coli str. K-12/MG1655. Similarly, for nfi, the highest-ranked target sequence was GATGTCGGGTTTGAGCAGGGCGG, chosen from 95 proposed sequences for this gene knockout in the same strain. Next, we used the first 20 nt of these sequences to design the gRNAs specific for ung and nfi, BBaK5061015 and BBaK5061016, respectively.

In E. coli, the DNA repair system for double-strand breaks is highly inefficient, rendering CRISPR/Cas9-induced breaks lethal to the cell [Widney et al., 2024]. Additionally, the native homologous recombination system is not effective too, therefore targeted repairs cannot be easily introduced without the use of phage recombination systems. To overcome this, the λ Red recombination system, derived from lambda (λ) phage, is commonly used as one of the most efficient tools for facilitating targeted genetic modifications [Datsenko and Wanner, 2000].

II.1.2.2 Build

Following the design of the two gRNA sequences specific for ung and nfi genes, we constructed two corresponding transcription units under the control of the J23119(SpeI) promoter (BBaK5061115 and BBaK5061116, respectively) that we assembled by CPEC (Circular Polymerase Extension Cloning) [Quan and Tian, 2011] in a backbone composed of the ampicillin resistance gene (AmpR) and the thermosensitive version of the oriR101 origin of replication.
In parallel, we amplified by PCR the ung and nfi repair templates (BBaK5061215 and BBaK5061216, respectively) using as template genomic DNA extracted from E. coli MG1655* Δflu ΔpyrF Δung Δnfi strain [Álvarez et al., 2020] that was kindly provided by Dr. Luis Ángel Fernández, at Centro Nacional de Biotecnología (CNB-CSIC) in Spain.

II.1.2.3 Test

E. coli S2060 cells were first transformed with the pCrepe plasmid carrying both the λ Red and Cas9 genes [Choudhury et al., 2020] and grown in LB media supplemented with 35 µg/mL chloramphenicol. Once the OD600 reaches 0.4, the culture was transferred to a water bath at 42°C for 15 minutes with shaking at 200 rpm to activate the expression of the λ Red recombination system proteins. The cells were then placed on ice for 15-20 minutes, then made electrocompetent. Subsequently, they were transformed by electroporation using 0.5 to 2 μg of plasmid DNA (expressing the gRNA) and 0.5 to 2 μg linear repair DNA in an electroporation cuvette at 1.8 kV and immediately resuspended in 1 mL of LB medium. After 3 hours of incubation at 30°C with shaking at 200 rpm, aliquots were spreaded on LB agar supplemented with chloramphenicol and ampicillin. The cultures were incubated overnight at 30°C. Isolated colonies were streaked on LB agar plates (without antibiotics) and cultured at 37°C to get cured from both plasmids (pCrepe and gRNA).

II.1.2.4 Learn

Unfortunately, the CRISPR/Cas9 and λ Red recombineering technique proved inefficient in our attempts to construct E. coli S2060 Δung Δnfi.
Although we have successfully assabled the gRNA expressing plasmids and readily obtained the repair templated by PCR, we were unable to obtain colonies upon their transformation in electrocompetent S2060 cells containing the pCrepe plasmid.
The process was particularly challenging due to the slow growth rate of the S2060 cells with the pCrepe plasmid, which had an estimated doubling time of 2 hours, making it difficult to align with the working hours of our host lab.
After multiple unsuccessful attempts, we opted to explore alternative methods for constructing E. coli S2060 Δung Δnfi.

II.1.3 Construction of E. coli S2060 Δung Δnfi through λ Red mediated recombineering

II.1.3.1 Design

As mentioned above, phage-derived recombination systems are commonly used to enable targeted genetic modifications in E. coli, with the λ Red system being one of the most effective tools. However, since recombination events are quite rare, a selection system is essential to identify successful modifications. The method developed by Datsenko and Wanner, 2000 addresses this need, and its effectiveness was demonstrated in the construction of the Keio gene deletion collection [Baba et al., 2006].

The method relies on the use of a linear donor DNA fragment containing 5′ and 3′ "homology arms" (H1/H2), matching the sequences immediately upstream and downstream of the target site. Upon expression of the λ Red genes carried on a thermosensitive plasmid like pKD46, the sequence between the homology arms in the donor DNA replaces the corresponding sequence in the target DNA. If this sequence includes an antibiotic resistance gene, successfully transformed cells become resistant to the antibiotic. Additionally, if the resistance gene is flanked by FRT sites, it can be excised by expressing the Flp recombinase from a second plasmid (pCP20 for instance [Cherepanov and Wackernagel]).

To use this method, we designed primers with H1 and H2 "homology arms" of 50 nt specific to the 5’ and 3’ sequences immediately upstream and downstream of both ung and nfi genes (Table 2).

Primer name Sequence (5'->3')
delta-ung-F TAGAAAGAAGCAGTTAAGCTAGGCGGATTGAAGATTCGCAGGAGAGCGAGagagcggccgccaccgcggg
delta-ung-R TGATAAATCAGCCGGGTGGCAACTCTGCCATCCGGCATTTCCCCGCAAATgcatatgctgcgtgcatgcg
delta-nfi-F TGGAGGCAGTGCATCGACTGTCTGAACAGTATCACCGCTAAGGAGTGATTagagcggccgccaccgcggg
delta-nfi-R TGTAACATGTTGAGTTCTCAAATACGGAAATTATCCGCAGTTTACCTGAAgcatatgctgcgtgcatgcg

Table 2. ung and nfi specific primers designed to amplify the Apramycin (ApraR) and Spectinomycin (SpecR) resistance genes carried on the pEVL408 [Trottier, 2019] and pEVL410 [Schulz, 2023] plasmids respectively. Nucleotides in capital letters are specific to the 5’ and 3’ sequences immediately upstream and downstream of ung or nfi genes, while those in lowercase letters are specific to the template DNA plasmids.

II.1.3.2 Build

No genetic constructs were necessary: all required plasmids were kindly provided by our host lab: pKD46 [Datsenko and Wanner, 2000], pCP20 [Cherepanov and Wackernagel], pEVL408 [Trottier, 2019] and pEVL410 [Schulz, 2023].
We (only) perform 2 PCR amplifications: one with the delta-ung-F and delta-ung-R (Table 2) on pEVL408 and another one with delta-nfi-F and delta-nfi-R (Table 2) on pEVL410.

II.1.3.3 Test

E. coli S2060 cells were first transformed with the pKD46 plasmid carrying the λ Red genes [Datsenko and Wanner, 2000] and grown in LB media supplemented with 100 µg/mL ampicillin. Once the OD600 reaches 0.6, the expression of the λ Red recombination system proteins was induced by adding L-arabinose to a final concentration of 10 mM. After 2 hours of incubation at 30°C with shaking at 200 rpm, cells were made electrocompetent and transformed with 0.5 µg PCR product amplified as described above. After 3 hours of incubation at 37°C with shaking at 200 rpm, aliquots were spreaded on LB agar supplemented with either 50 µg/mL aparamycin or 100 µg/mL spectinomycin (depending on which plasmid pEVL408 or pEVL410 respectively was used as template for the PCR reaction). The cultures were incubated overnight at 37°C to eliminate the pKD46 plasmid.
Δung ApraR colonies were selected on apramycin and the correct insertion confirmed by colony PCR with external primers binding the grcA gene (GAACTTCTGGTTTCACTTCTACTGGAACTTCACG) and yfiF gene (CAGGATTGCGTTCTGGCACTGG).
Δnfi SpecR colonies were selected on apramycin and the correct insertion confirmed by colony PCR with external primers binding the hemE gene (GCGTCACTTTGTATCTGAATGCGCAGATTAAAGC) and yjaG gene (CTCATTTCGCGACCAGCCTGAGTCATTTCC).
In a subsequent step, antibiotic resistant cells were transformed with the pCP20 plasmid and selected on LB agar plates supplemented with 100 µg/mL carbenicillin. An isolated colony was then grown at 42°C to activate the expression of the Flp recombinase, then streaked on LB agar places containing no antibiotic. The pCP20 plasmid was eliminated by growing the cells at 37°C.
The correct excision of the antibiotic was confirmed first by colony PCR with the external primers mentioned above, then by Sanger sequencing of the PCR product at Eurofins Genomics.

Finally, genomic DNA of the E. coli S2060 Δung Δnfi strain was extracted using the PureLink™ Genomic DNA Mini Kit (Invitrogen) from 1 mL overnight culture following the supplier’s Gram Negative Bacterial Cell Lysate Protocol. The concentration of dsDNA was evaluated with the Qubit™ 1X dsDNA BR (Broad Range) Assay (Invitrogen), then sequenced at Eurofins Genomics using Oxford Nanopore Technologies for bacterial genome sequencing.

II.1.3.4 Learn

λ Red recombineering technique following the protocol described by Datsenko and Wanner, 2000 proved very efficient and we have successfully constructed the E. coli S2060 Δung Δnfi.
It was a multi step process in which we first inserted the ApraR resistance cassette at the ung locus, then used the Flp recombinase to excise it. Afterwards, we inserted the SpecR resistance cassette at the nfi locus and finally we used the Flp recombinase to excise it too.
At each step, several clones were isolated through multiple passages on appropriate antibiotics and analyzed by colony PCR and Sanger sequencing. Regularly, antibiotic resistance profiles were checked on many antibiotics in parallel, to avoid any carry over of the ‘previous’ stain and sudden reappearance of the wild type phenotype.

Finally, we used the bacterial genome sequencing service provided by Eurofins Genomics that employs state-of-the-art Oxford Nanopore Technologies long-read sequencing technology, and completely sequenced the genome of our E. coli S2060 Δung Δnfi strain.

The first analysis of the NGS data was performed by Eurofins Genomics, using their established pipeline:
“Raw nanopore sequencing reads are assessed for quality and filtering. Filtlong v0.2.1 [https://github.com/rrwick/Filtlong] is used to remove short and low-quality reads from the raw nanopore sequencing data. The high-quality nanopore sequencing reads are then used for de novo assembly of the bacterial genome. The assembly is performed using Flye v2.9.3 [Kolmogorov et al., 2019] with parameters optimized for bacterial genomes. The resulting contigs are further polished using Medaka v1.8 [https://github.com/nanoporetech/medaka] to improve base accuracy. The assembled genomes are annotated using Bakta v.1.8.2 [Schwengers et al., 2021]. The annotation includes the prediction of coding sequences, tRNAs, rRNAs, and other genomic features based on published databases such as RefSeq and UniProt. The quality of the assembled genomes is assessed using various quality assessment tools (QUAST v5.2 [Gurevich et al., 2013], CheckM2 v1.0.1 [Chklovski et al., 2023], Mash v2.3 [Ondov et al., 2016]). Genome completeness, contiguity, and accuracy are evaluated to ensure the reliability of the assemblies. A final QC step is included to ensure the purity of samples, which uses minimap2 v2.24 [Li, 2018] to map the sequence-cleaned reads onto the assembly and finally employs Clair3 v1.0.4 [Zheng et al., 2022] to call variations (SNPs and INDELS) within the assembled genome. Any variants detected along with the location in the bacterial assembly are reported.”

Through this professional analysis, 6 contigs were assembled for a total of 4827035 nucleotides (nt), with an average coverage depth of the assembled genome of 42 and an overall GC content of 50.67%. The machine learning algorithm used to assess the assembly completeness estimated it at 100% and revealed a 1.78% contamination rate. Importantly, no variants were detected in the ‘Purity Check’ of the assembled genome during which the sequenced reads are mapped against the assembled genome and the variants (SNPs, insertions & deletions) are determined and reported if having at least 0.3 minor allele frequency(MAF) and >30x read support.

The 6 contigs assembled by Eurofins Genomics have, in descendent order: 4168152 nt, 266411 nt, 198664 nt, 112058 nt, 79301 nt and 2449 nt.

We performed further sequence analysis and compared them with the published genome of E. coli DH10B [Durfee et al., 2008], the parental strain of S2060 (Table 1).
The longest contig of 4168152 nt really aligned on DH10B (GenBank NC_010473), albeit with a 507224 nt gap at position 233634.. 740857. Contigs 2, 4, 5 and 6 partially covered this gap as they all aligned in this region (Figure 9). It is to be noted that contig 6, the smallest one carries the proA and proB genes that in S2060 are no longer in the chromosome, but on the F’ plasmid.

Contig 3, of 198664 nt, only partially and to a very limited extent aligned to the DH10B genome. A blast analysis revealed that it’s the F’ plasmid, with the F'Iq (GenBank CP053608) being the top best hit (Score 2.968e+05, 95% query cover and 100% identity over 160718 nt). The F’ plasmid of S2060 was also engineered starting from the one carried by E. coli ER2738, it is thus not unexpected to observe differences.

This comparative genomics analysis allowed us to investigate and map the genetic differences compared to DH10 and thus confirm the genetic modifications made to the S2060 strain. The alignments confirmed that the ung and nfi genes are no longer present in our strain and that the sequence is as expected and exactly the same as the one Sanger sequencing of the colony PCR products.


Figure 9. Alignment of the contigs assembled by Eurofins Genomics on the genome of E. coli DH10B. The alignment was generated using SnapGene (to circumvent le 1000000 nt limit, the largest contig was splitted in 5, then the image was edited to remove the artifacts). Mutations indicated on top are expected according to the strain genotype, while those indicated in the bottom were unexpected and mapped following the breseq analysis.

Moreover, to understand why this gap was present in the assembled contigs by the sequencing company, we performed further analysis using minimap2 [Li, 2018, 2021] and breseq [​​Deatherage and Barrick, 2014].

Minimap2 quickly showed that Nanopore reads were aligned in the gap region from position 233.634 to 740.857 on DH10B genome (Figure 10). The alignments in this part showed differences in depth of coverage within the gap, but also with the rest of the alignment outside this region. This may explain the failure of de novo assembly algorithms to properly assemble this region in a contig that meets the quality standards.


Figure 10. Alignment of the Nanopore reads on the genome of E. coli DH10B with a zoom on the 233,634 .. 740,857 position. The alignment was generated using minimap2 and the visualization was made with Integrative Genomics Viewer (IGV, v2.18.2) [Robinson et al., 2011].

The analysis performed with breseq allowed us to map and identify the differences between S2060 and DH10B. The full results are available HERE Notably, no reads mapped the ung and nfi locus which undoubtedly confirmed thor deletion (Figures 11 and 12).


Figure 11. Alignment of the Nanopore reads on the genome of E. coli DH10B with a zoom on the ung locus. The alignment was generated using breseq.


Figure 12. Alignment of the Nanopore reads on the genome of E. coli DH10B with a zoom on the nfi locus. The alignment was generated using breseq.

II.2 Selection Phages (SP) for XylS evolution

In PACE / PANCE / PRANCE and PHAGEVO, the production of selection phages (SP) is a preliminary, but essential step without which evolution cannot be started.

II.2.1 Design

As described above, the SP is a modified M13 phage lacking a gene responsible for its infectivity which is replaced by the gene of interest to be evolved (GOI).
Traditionally, it's the gIII that was chosen as a selection marker [Esvelt et al., 2011], but, as its low expression makes the host cell resistant to infection [Boeke et al., 1982] which hinders phage auto propagation and the auto screening of the best variants of GOI, gVI was developed as an alternative [Brödel et al., 2016; Miller et al., 2020b]. Based on this, in our setup we choose to implement in PHAGEVO two different approaches: one using gVI and the other using both gIII and gVI as selection markers.
Genetically engineering phages is not as easy as for plasmids. For this reason, helper phages (HP) were developed. M13KO7 is one known example of such HP that includes the origin of replication from p15A and a kanamycin resistance gene, both inserted into M13 origin of replication [Vieira and Messing, 1987]. The presence of these two elements allows it to replicate as a plasmid in E. coli cells, but does not prevent it to replicate, pack and secrete single-stranded phage DNA. M13KO7 is commercially available from New England Biolabs (Cat#N0315S), and we decided to use it as a platform for the PHAGEVO HP constructs.
To allow the evolution of XylS, we designed a total of 10 helper phages (Table 3), two of which are Golden Gate platforms to facilitate the insertion of the GOI and eight of them are XylS-specific HP. Our set of HP is composed of half of HP designed for using gVI as a selection marker and the other half for the gIII-gVI.
The XylS-specific HP are of several types:

In PANCE, the GOI is equipped with an RBS and inserted in the HP in place of the selection marker (gIII and/or gVI). This disrupts the expression of the downstream gI gene in the M13 genome, thus the GOI is followed by the J23107 constitutive promoter and an RBS.
In PHAGEVO, we keep this general architecture and added upstream and downstream of our GOI the T7 promoters and terminators as illustrated in Figure 8.

Insertion of the GOI in the HP leads to the disruption of the infectious cycle of the M13 phage as at least an important gene in this process is missing. To complete this cycle, its expression in trans from the accessory plasmid (AP) is required.
In our project we designed a total of 6 AP (Table 4):

These APs express in an operon not only the missing gene from the HP (gVI or gIII-gVI), but also the LuxAB, which is a fusion protein between LuxA and LuxB from Photorhabdus luminescens (Xenorhabdus luminescens) via a 21 amino acids flexible linker. LuxA, the alkanal monooxygenase alpha chain (Uniprot P19839) is a 360 amino acids protein involved in light emission by luminous bacteria together with LuxB, the alkanal monooxygenase beta chain of 327 amino acids (Uniprot P19840) [Johnston et al., 1990, Xi et al., 1991]. The LuxAB complex is an FMN bound enzyme (EC:1.14.14.3) catalyzing the oxidation of an aliphatic long-chain-aldehyde to the corresponding fatty acid. The enzyme exhibits high specificity for reduced FMN and long-chain aliphatic aldehydes containing eight or more carbon atoms, with maximum efficiency observed using tetradecanal (aka myristyl aldehyde) [Li and Meighen, 1994]. The aldehyde is synthesized in vivo by the action of the LuxCDE operon, expressed from the F’ plasmid carried by the host cells of E. coli S2060 strain, as described above. The expression of luciferase (LuxABCDE) allows monitoring the phage propagation through luminescence monitoring experiments [Badran et al., 2016; DeBenedictis et al., 2022].

Table 3. Helper phages (HP) designed and built for the evolution of XylS through PANCE and PHAGEVO.

Usage HP Description HP Part Number
Golden Gate cloning platform Helper phage (HP) M13KO7-ΔgVI Golden Gate cloning platform BBa_K5061031
Golden Gate cloning platform Helper phage (HP) M13KO7-ΔgIII-ΔgVI Golden Gate cloning platform BBa_K5061131
PACE / PANCE Helper phage (HP) M13KO7-ΔgVI carrying XylS K38R L224Q BBa_K5061051
PACE / PANCE Helper phage (HP) M13KO7-ΔgIII-ΔgVI carrying XylS K38R L224Q BBa_K5061151
PHAGEVO Helper phage (HP) M13KO7-ΔgVI carrying XylS K38R L224Q in the PHAGEVO mutational region BBa_K5061052
PHAGEVO Helper phage (HP) M13KO7-ΔgIII-ΔgVI carrying XylS K38R L224Q in the PHAGEVO mutational region BBa_K5061152
PACE / PANCE Helper phage (HP) M13KO7-ΔgVI carrying XylS W88C L224Q BBa_K5061053
PACE / PANCE Helper phage (HP) M13KO7-ΔgIII-ΔgVI carrying XylS W88C L224Q BBa_K5061153
PHAGEVO Helper phage (HP) M13KO7-ΔgVI carrying XylS W88C L224Q in the PHAGEVO mutational region BBa_K5061054
PHAGEVO Helper phage (HP) M13KO7-ΔgIII-ΔgVI carrying XylS W88C L224Q in the PHAGEVO mutational region BBa_K5061154

Table 4. Helper phages (HP) designed and built for the evolution of XylS through PANCE and PHAGEVO.

Usage AP Description AP Part Number
Golden Gate cloning platform Accessory Plasmid (AP) Golden Gate cloning platform for M13 gVI expression BBa_K5061032
Golden Gate cloning platform Accessory Plasmid (AP) Golden Gate cloning platform for M13 gIII and gVI expression BBa_K5061132
activity-independent phage growth Accessory Plasmid (AP) expressing the M13 gVI and LuxAB from the PspA promoter BBa_K5061045
activity-independent phage growth Accessory Plasmid (AP) expressing the M13 gIII, M13 gVI and LuxAB from the PspA promoter BBa_K5061145
XylS-dependent phage growth Accessory Plasmid (AP) expressing the M13 gVI and LuxAB from the Pm promoter BBa_K5061055
XylS-dependent phage growth Accessory Plasmid (AP) expressing the M13 gIII, M13 gVI and LuxAB from the Pm promoter BBa_K5061155

II.2.2 Build

All HP and AP plasmids (Tables 3 and 4) were assembled by Golden Gate using either BsaI or BsmBI type IIS restriction enzymes.
First we constructed the HP and AP Golden Gate cloning platforms from DNA fragments obtained either by PCR from plasmid templates purchased from New England Biolabs (M13KO7) or from Addgene (pJC175e) or synthesized. During this process, care was taken to remove BsaI and BsmBI sites from M13KO7 and pJC175e. Upon sequencing, differences compared to the M13KO7 HP sequence available on New England Biolabs website were revealed, the reason for which we added this M13KO7 to the Parts Registry (BBa_K5061000). Some of these differences are present in another M13 helper phage VCSM13 (GenBank AY598820). A complete list is presented on the Parts Registry page.
Using these ‘universal’ Golden Gate cloning platforms for HP and AP, we readily built all the other HP and AP plasmids.

II.2.3 Test

Selection phage (SP) production was carried out in E. coli S2060 cells in three different conditions in a XylS-dependent or independent manner, with or without antibiotics selection as described by Miller et al., 2020. The thus produced phages were analyzed by plaque assay and luminescence monitoring.

XylS-independent SP production without antibiotics selection
The E. coli S2060 cells were first transformed with the Accessory Plasmid (AP) expressing either the M13 gVI or gIII-gVI and LuxAB from the PspA promoter (BBaK5061045 and BBaK5061145, respectively). Transformed cells were made competent and then transformed with the various HP constructs (Table 3). Care was taken for matching the selection markers (gVI and gIII-gVI) between the HP and the AP. As controls, the HP Golden Gate cloning platform as well as the M13KO7 were included. At the end of the transformation protocol, cells were cultured in 10 mL liquid LB media (without any antibiotic) at 37°C with shaking at 200 rpm. After this overnight incubation, cells were pelleted by centrifugation (4000 rpm for 10 minutes at 4°C), and the supernatant filtered using a 10 mL syringe fitted with a 13-mm, 0.22 µm PVDF or PES syringe filter to remove residual cells. Filtered phages were stored at 4°C.

XylS-independent SP production with antibiotics selection
As the transformation procedure is not always very efficient especially when using home-made prepared competent cells, we modified the above described protocol and cultured the transformed cells in LB media supplemented with 5 µg/mL Tetracycline (to select for F’ containing S2060 cells), 12.5 µL kanamycin (to select for HP containing cells) and 50 µg/ml ampicillin (to select for AP containing cells).

XylS-dependent SP production
The XylS-dependent SP production was carried out slightly differently depending on whether the PANCE of PHAGEVO HP were used:

These E. coli cells were then co-transformed with the various HP constructs (Table 3) and the corresponding AP expressing either the M13 gVI or gIII-gVI and LuxAB from the Pm promoter (BBaK5061055 and BBaK5061155, respectively). Care was taken for matching the selection markers (gVI and gIII-gVI) between the HP and the AP. As controls, the HP Golden Gate cloning platform as well as the M13KO7 were included. At the end of the transformation protocol, cells were cultured in 10 mL liquid LB media supplemented with 1 mM m-toluate to activate XylS that induces the expression of the M13 phage protein expressed from the AP plasmid. After an overnight incubation at 37°C with shaking at 200 rpm, cells were pelleted by centrifugation (4000 rpm for 10 minutes at 4°C), and the supernatant filtered using a 10 mL syringe fitted with a 13-mm, 0.22 µm PVDF or PES syringe filter to remove residual cells. Filtered phages were stored at 4°C.

Plaque assay
Plaque assay were performed in an activity-independent manner using E. coli S2060 cells carrying the AP expressing either the M13 gVI or gIII-gVI and LuxAB from the PspA promoter (BBaK5061045 and BBaK5061145, respectively).
For this 150 µL of these recipient cells at an OD600 between 0.6 and 1 were mixed with 10 µL of SP phages prepared as described above, 10 mL of liquid Soft LB Agar (10 g/L Tryptone, 5 g/L Yeast Extract, 5 g/L NaCl, 7.5 g/L Agar, melted and cooled down at 55°C), and 40 µL of Bluo-Gal 20 mg/mL, then rapidly spread on LBagar plates (without antibiotics). After an overnight incubation at 37°C, lysis plaques were observed with the naked eye.

Luminescence monitoring
Luminescence monitoring of phage propagation was performed in an activity-independent manner using E. coli S2060 cells carrying the AP expressing either the M13 gVI or gIII-gVI and LuxAB from the PspA promoter (BBaK5061045 and BBaK5061145, respectively).
For this 10 µL of these recipient cells at an OD600 between 0.6 and 1 were mixed with 10 µL of SP phages prepared as described above in 200 µL of LB medium supplemented with 5 µg/mL Tetracycline (to select for F’ containing S2060 cells) and 50 µg/mL ampicillin (to select for AP containing cells) and 400 ng/µL anhydrotetracycline in an opaque 96-well polystyrene microplate (COSTAR 96, Corning). The plate was then incubated overnight at 37°C with shaking at 200 rpm and the luminescence and optical density at 600 nm (OD600) were taken every 15 minutes using a CLARIOstar (BMGLabtech) plate reader. Luminescence values were normalized by OD600 to account for variations in cell density.

II.2.4 Learn

Both luminescence and plaque assays demonstrated the successful production of selection phages (SP) in both XylS-dependent and independent manner (Figures 12, 13 and 14).

Luminescence monitoring in 96-well plate proved to be an easy way to monitor phage propagation, the reason for which the E. coli S1030 strain was initially developed [Carlson et al., 2014]. Typical bacterial growth curves and the luminescence output are presented in Figure 12, where we can observe a shift between the 2 curves indicating that the phages are produced mainly in the stationary growth phase.
Using this approach, we evaluated the SP phages produced in three different conditions in a XylS-dependent or independent manner, with or without antibiotics selection, along with the controls (Figure 13). When available, we tested several clones obtained upon HP assembly by Golden Gate.
Analyzing the results presented in figure Figure 13, we can observe that high Luminescence/OD600 values were obtained as expected for the positive controls performed with the M13KO7 taken directly from the tube purchased from New England Biolabs (NEB) or from several preparations performed by us starting from this commercial HP.
When phages were prepared in an activity-independent manner (Figures 13A and 13B), we can observe variable degrees of Luminescence/OD600 values indicating efficient and less efficient phage preparations or testing conditions. Transformation efficiency during phage preparation may account for these variations which may lead to higher or lower phage titer. Testing conditions are also important, the output being dependent on the proper complementation by the AP of the missing phage protein or the differences in cell densities of the host cells. Indeed, we can observe that lower Luminescence/OD600 values were obtained in general when the gIII-gVI selection marker was used compared to gVI alone, including for the M13KO7 positive control. This experiment was carried out in parallel for all samples, with two different host cells, an inherent variable to this type of experiment.
Variable degrees of Luminescence/OD600 values were also observed when phages were prepared in a XylS activity-dependent manner (Figure 13C). In this case, XylS production and induction of Pm promoter may be responsible for these differences. It is remarkable the highest Luminescence/OD600 value was obtained with a phage prepared in a XylS activity-dependent manner, demonstrating the efficiency of our experiment setup.


Figure 12. In vivo characterization of the Selection Phage (SP) propagation in E. coli S2060 cells carrying an Accessory Plasmid (AP) expressing LuxAB from the PspA promoter. Example of a growth curve (A) and of the luminescence produced during phage multiplication (B) and the corresponding Luminescence/OD600 values (C).


Figure 13. In vivo characterization of our various Selection Phage (SP) preparation through luminescence monitoring in E. coli S2060 cells carrying an Accessory Plasmid (AP) expressing LuxAB from the PspA promoter. (A) SP prepared in an activity-independent manner without antibiotics selection. (B) SP prepared in an activity-independent manner with antibiotics selection. (C) SP prepared in a XylS activity-dependent manner. PANCE and PHAGEVO denote the absence of presence of T7 promoters and terminators flanking the XylS gene. SP phages carry either the XylS-K38R-L224Q or the XylS-W88C-L224Q, the two variants reported in the literature as being able to weakly detect PA and TPA, respectively [Li et al., 2022].


Figure 14. In vivo characterization of the Selection Phage (SP) propagation in E. coli S2060 cells carrying an Accessory Plasmid (AP) expressing LuxAB from the PspA promoter. Example of lysis plaques obtained with two different SP preparations.

II.3 XylS evolution: experimental setup and phage library analysis

In PACE / PANCE / PRANCE and PHAGEVO, the evolution of the GOI happens in a reaction vessel called lagoon, where phages infect the bacteria, replicate, and then go on to infect new bacterial cells, as illustrated in Figures 4 and 8. Setting up this evolution involves mixing the selection phages with the host cells, applying a selection pressure, monitoring the phage propagation before harvesting them and analyzing the activity of the GOI.

II.3.1 Design

Performing XylS evolution experiments required not only the preparation of selection phages described above, but also the preparation of the appropriate lagoon host cells.
Selection phages are lacking either the gVI or both gIII and gVI. As for the experiments performed during the phage production, host cells need to carry the AP plasmids expressing these missing genes. We used the same AP plasmids listed above (Table 4) that allowed us, when using the Pm promoter ones, to apply a selection pressure depending on the level of XylS induction by either PA or TPA. The existing AP plasmids also allowed us to have positive and negative controls over phage infection when the PspA phage shock or no promoter conditions were used respectively. Moreover, these AP plasmids allow real-time monitoring of phage propagation by luminescence in E. coli S2060.
In addition, to increase mutation rates, host bacterial cells carry a mutagenesis plasmid (MP in Figures 4 and 8). The most potent one described in the literature is MP6 which expresses 6 genes (dnaQ926, dam, seqA, emrR, ugi and PmCDA1) that insert mutations, block the DNA repair system and are involved in the DNA methylation system [Badran and Liu, 2015]. The drawback of this system is that mutations are inserted all over the genome and not only on our gene of interest, which can result in loss of interesting variants if the system becomes non-functional due to mutations in essential genes.
In contrast, in Evolution.T7 [https://2021.igem.org/Team:Evry_Paris-Saclay], mutations are focused in the GOI, thanks to the specificity of the T7 RNA polymerase (T7RNAP) for its cognate promoter and terminator flanking the GOI as illustrated in Figure 5. The Evolution.T7 constructs expressing the T7RNAP linked to a base deaminase (BD) either a cytosine or an adenosine deaminase (respectively CD and de notre ère) are carried by low copy plasmids (pSEVA221 and pSEVA471) and the expression is inducible by anhydrotetracycline for BD-T7RNAP and by L-arabinose for BD-T7RNAPCGG-R12-KIRV.
In PHAGEVO we decided to construct 4 mutagenic plasmids to express the most effective of the CD and de notre ère deaminases of Evolution.T7, evoCDA1-BE4max and ABE8.20-m respectively, each fused to the T7RNAP and T7RNAPCGG-R12-KIRV for forward and reverse sense respectively. These constructs aimed also to harmonize the expression system by using the same pBad promoter and the same backbone as the MP6 plasmid. The list of all mutagenic plasmids and their controls is presented in Table 5.

Table 5. Mutagenesis plasmids (MP) designed, built or purchased and used for the evolution of XylS through PANCE and PHAGEVO.

As a bacterial host cell, we decided to construct the E. coli S2060 ∆ung ∆nfi that combines the advantages of the PACE strain (the S2060) with those of the Evolution.T7 strain (∆ung ∆nfi), as described above.

Several host cells carrying combinations of various AP and MP plasmids are thus possible. Considering also the number of selection phages we prepared and tested (Figure 13), the number of XylS evolution experiments to be conducted is not a negligible one. This combinatorial explosion was noticed early on in the development of our project, and, to be able to test many conditions in parallel, we decided to adopt the non continuous PANCE strategy to implement the PHAGEVO system.

II.3.2 Build

All MP plasmids (Table 5) were assembled by classical restriction digestion between MP6 and the Evolution.T7 plasmids build by the iGEM Evry Paris-Saclay 2021 team (BBaK3766109, BBaK3766110, BBaK3766106, BBaK3766107).
The EP Golden Gate cloning platform, used here as a negative control as it is an empty backbone with the same origin of replication and the same antibiotic resistance marker as the MP construct, was assembled by Golden Gate using the BsmBI type IIS restriction enzymes and DNA fragments obtained either by PCR from plasmid templates purchased from Addgene (MP6) or synthesized.
The construction of AP plasmids is presented above, as well as that of the E. coli S2060 ∆ung ∆nfi strain.

II.3.3 Test

XylS evolution experiments
XylS evolution experiments were performed in 96-well microplates and were monitored through luminescence measures as described above.
The E. coli S2060 ∆ung ∆nfi cells were first co-transformed with combinations of the various AP and MP plasmids listed in Tables 4 and 5, respectively. Transformed cells were selected and propagated in LB media supplemented with 5 µg/mL tetracycline (to select for F’ containing S2060 ∆ung ∆nfi cells), 50 µg/ml ampicillin (to select for AP containing cells), 17.5 µL chloramphenicol (to select for MP containing cells) and 100 mM Glucose to catabolically inhibit any potential leaky expression form the pBad promoter and thus the expression of the mutagenic genes.
Then, 10 µL of these recipient cells at an OD600 between 0.6 and 1 were mixed with 10 µL of SP phages prepared as described above in 200 µL of LB medium supplemented with 5 µg/mL tetracycline, 50 µg/mL ampicillin, 17.5 µg/mL chloramphenicol, 400 ng/µL anhydrotetracycline, 0.2% L-arabinose (to induce the expression of the mutagenic genes from the pBad promoter) and 100 µM of either PA or TPA. Control experiments without PA / TPA were also performed.
The opaque 96-well polystyrene microplate (COSTAR 96, Corning) was then incubated overnight at 37°C with shaking at 200 rpm and the luminescence and optical density at 600 nm (OD600) were taken every 15 minutes using a CLARIOstar (BMGLabtech) plate reader. Luminescence values were normalized by OD600 to account for variations in cell density.
Cultures displaying Luminescence/OD600 values higher in the presence of MP plasmids compared to the values obtained in their absence were selected, transferred in 1.5 mL microtubes and centrifuged (7000 rpm for 3 minutes at 4°C) and the supernatant filtered using a MultiScreen-GV Sterile, clear 96-well filter plate with 0.22 um pore size Hydrophilic PVDF membrane (Millipore) fixed on a DeepWell plate. Filtered phages were stored at 4°C.

Analysis of evolved phages
Evolved phages were analyzed by plaque assay and luminescence monitoring following the same protocols as described above.
As their volume is low ( <200 µL), an aliquot was used to re-infect host cells and were thus multiplied in a XylS-independent manner as described above.
Phage DNA was prepared using the E.Z.N.A. M13 DNA Kit (Omega Bio-Tek) and analyzed by spectrophotometry.
To produce dsDNA for sequencing purposes, we used the Illustra TemphiPhi amplification kit (Cytiva) and performed rolling circle amplification (RCA) on 0.5 µL of ether the evolved phages directly recovered from the evolution 96-well plate, or on the phage DNA extracted with the E.Z.N.A. M13 DNA Kit, following the supplier's protocol. RCA products were further digested with the AvrII/XmaJI restriction enzyme, then purified using the Monarch DNA CleanUp kit (New England Biolabs). The concentration of dsDNA was evaluated with the Qubit™ 1X dsDNA BR (Broad Range) Assay (Invitrogen), then the samples having a concentration above 30 ng/µL were sequenced at Eurofins Genomics using Oxford Nanopore Technologies for clonal / linear amplicons.

Evolved phages were also used as templates to amplify the XylS coding sequence as detailed on the Parts Collection page on this wiki.

II.3.4 Learn

96 XylS evolution experiments were performed in parallel in 96-well plates using all phage preparations described above. As the number of possible combinations between the SP, AP, MP and PA/TPA inducers was high, we only assessed MP6 as a mutagenic plasmid for SP phages having the XylS gene not flanked by the T7 promoters and terminators (classical PANCE), and all 5 mutagenic plasmids on constructs having them (PHAGEVO). Moreover, evolution experiments toward PA detection were initiated only with SP phages carrying the XylS-K38R-L224Q gene while the evolution toward TPA detection were initiated only with SP phages carrying the XylS-W88C-L224Q gene as these variants were reported to have the ability to weakly detect the plastic degradation products PA and TPA, respectively [Li et al., 2022].
Luminescence monitoring of these evolution experiments allowed us to identify 30 phage populations showing increased Luminescence/OD600 values in the presence of a given MP plasmids compared to the values obtained with no mutagen (Figure 15).

These 30 phage populations were isolated and the presence of phages was confirmed by plaques assays (data not shown) and through luminescence monitoring (Figure 16). Data show that some phages are present in higher titer than others in the various preps.


Figure 15. Luminescence monitoring of in vivo evolution experiments in E. coli S2060 ∆ung ∆nfi cells carrying an Accessory Plasmid (AP) expressing LuxAB from the Pm promoter, various mutagenic plasmids and (A) SP prepared in an activity-independent manner without antibiotics selection, (B) SP prepared in an activity-independent manner with antibiotics selection, (C) SP prepared in a XylS activity-dependent manner. PANCE and PHAGEVO denote the absence or presence of T7 promoters and terminators flanking the XylS gene, respectively. SP phages carry either the XylS-K38R-L224Q or the XylS-W88C-L224Q, the two variants reported in the literature as being able to weakly detect PA and TPA, respectively [Li et al., 2022]. Data represent the ratio between the Luminescence/OD600 values observed in the presence of a given MP plasmid compared to the values obtained with no mutagen.


Figure 16. In vivo characterization of the 30 selected Evolved Phages through luminescence monitoring in E. coli S2060 cells carrying an Accessory Plasmid (AP) expressing LuxAB from the PspA promoter. PANCE and PHAGEVO denote the absence or presence of T7 promoters and terminators flanking the XylS gene. PA and TPA denote the target compound for which the phages were evolved for starting from XylS-K38R-L224Q or XylS-W88C-L224Q, respectively.

To further characterize them, M13 ‘minipreps’ were performed and the single stranded DNA (ssDNA) amount was estimated by spectrophotometry. The results presented in Figure 17 show that the amount of purified ssDNA is around 10 ng/µL (for an elution volume of 50 µL). The 260/280 ratios are, for the majority of the samples, around 1.8, a value generally accepted as an indication of ‘pure’ for DNA. However, the 260/230 ratios are low, indicating the presence of contaminants which absorb at 230 nm like EDTA, guanidine, trizol, phenol, carbohydrates.


Figure 17. Spectrophotometric analysis on M13 ssDNA preparations of evolved phages. (A) ssDNA concentration estimated based on the absorbance values at 260 nm and an extension coefficient of 33 ng/µL cm-1 (B) 260/280 and 260/230 Ratios.

Furthermore, in order to analyze the sequence variability of the evolved phages, the phage ssDNA was subject to rolling circle amplification (RCA), and the thus obtained produced digested with AvrII/XmaJI restriction enzyme (that has a unique site in all SP in the J23107 promoter driving the expression the gI M13 gene).
3 samples only had a dsDNA concentration compatible with the Eurofins Genomics requirements for NGS sequencing using Oxford Nanopore Technologies for clonal / linear amplicons. As of today results are still pending, but we’ll be happy to present them at the Jamboree.

Finally, the 30 evolved phage preparations were used in PCR reactions that readily allowed the amplification of the XylS gene with specific primers. All the experimental details and the subsequent results obtained for this experiment are detailed in the "Parts Collection" page on this wiki.

CONCLUSIONS

The goal of this year iGEM Evry Paris-Saclay team was to develop a new tool for in vivo continuous evolution of proteins, with improved properties compared to previous technologies. We came up with this idea and decided to create PHAGEVO: a new system that takes advantage of the best aspects of two powerful directed evolution technologies: the Phage-Assisted (Non) Continuous Evolution, or PA(N)CE, and Evolution.T7. The former has the big advantage to continuously evolve and select, in the same reaction medium, for the variants of the gene of interest with improved activity. The latter is a targeted in vivo mutagenesis system based on a modified T7 RNA polymerase fused with base deaminase (BD-T7RNAP) that mutate a region of the genome flanked on both 3’ and 5’ end by specific promoters of the BD-T7RNAP fusions, enabling mutation on both DNA strands.

Our team successfully designed and implemented the new PHAGEVO system. For this purpose, we first engineered the E. coli S2060 ∆ung ∆nfi strain, also known as the PHAGEVO E. coli strain, and confirmed its genotype by whole bacterial sequencing using Nanopore technology. Then, we engineered bacteriophage M13 to insert our gene of interest, XylS with the Evolution.T7 cassette in the phage genome. We also constructed several different plasmids constructs for both mutagenesis (mutagenesis plasmids) and selection (accessory plasmids) based on Evolution.T7 and latest improvements of the PACE technology. Different combinations of these plasmids where used in the subsequent evolution experiment.

We used the newly developed PHAGEVO system to run evolution experiments on the promiscuous benzoate derivative sensitive XylS transcription factor to make it more sensitive and specific toward either Phthalic Acid (PA) or Terephthalic Acid (TPA). Phage production in both XylS-dependent and independent manner was achieved, demonstrating the complementation between XylS and its cognate promoter in the PHAGEVO setup.

Finally, evolved phages were used to isolate a collection of XylS variants that were further analyzed as described on the Parts Collection page on this wiki.

REFERENCES