Strains, plasmids, media
All assay experiments were performed in the Escherichia coli strain 3.32 (lacZ13(Oc) lacI22 λ−el4- relA1 spoT thiE1; Yale CGSC #5237), transformed chemically, while standard DNA cloning was performed in NEB 5-alpha Competent E. coli (huA2 Δ(argF−lacZ)U169 phoA glnV44 φ80Δ(lacZ)M15 gyrA96 recA1 relA1 endA1 thi-1 hsdR17; New England Biolabs) and library construction was performed in electrocompetent E. coli 3.32 cells. Cells were grown in LB Miller Medium (Fisher Scientific) or M9 Minimal Medium (6.8 g L−1 Na2HPO4, 3.0 g L−1 KH2PO4, 0.5 g L−1 NaCl, 1.0 g L−1 NH4Cl, 2 mM MgSO4, 100 μM CaCl2; Millipore Sigma) supplemented with 0.2% (w/v) casamino acids (VWR Life Sciences), 1 mM thiamine HCl (Alfa Aesar). LB Miller + Agar (Fisher Scientific) was used for selection when cloning. Antibiotics and ligands were used as appropriate. Antibiotics used were: chloramphenicol (25 μg mL−1; VWR Life Sciences), kanamycin (35 μg mL−1; VWR Life Sciences), and carbenicillin (100 μg mL−1; Teknova). Ligands used were: fructose (Arcos Organics), d-ribose (Arcos Organics), and isopropyl-β-d-thiogalactoside (IPTG; Millipore Sigma). All ligands in this study were used at a concentration of 10 mM.
Cloning and transcription factor plasmid construction
For all cloning experiments, oligo synthesis and DNA sequencing were performed by Eurofins Genomics. All DNA constructs (genetic architectures and protein mutants) were sequenced to verify mutations, sequence identity, and correct assembly following transformation and plasmid isolation (Omega Bio-Tek). Sequence alignments and primer design were aided by ApE Plasmid Editor and SnapGene software. All polymerase chain reactions (PCR) were performed using Phusion High-Fidelity PCR Master Mix with GC Buffer (NEB), except for the directed evolution library, which used Taq DNA Polymerase (NEB). Plasmids with chimeric transcription factors in the pLacI (Novagen) backbone (featuring a low copy number p15A origin, chloramphenicol resistance marker, and the constitutive LacI promoter driving a transcription factor gene) were taken from Rondon, et al.25. Briefly, these were constructed by obtaining the respective open reading frames (ORFs), RbsR-L (Addgene #60773), and FruR-L (Addgene #60768), gifts from the Swint-Kruse and Bennett Labs, and inserting them into pLacI using circular polymerase extension cloning (CPEC)52 following amplification of the genes. Likewise, pLacI vectors with wild-type FruR (UniProt # P0ACP1) and RbsR (UniProt #P0ACQ0) ORFs were constructed by inserting a genomic copy (guided by Blattner et al.53) of the ORFs into pLacI using CPEC following amplification of the gene via colony PCR from E. coli 3.32 cells. As done by Rondon et al.25, FruR-L was driven by the LacIq (rather than LacI) promoter54, which leads to approximately a ten-fold increase in protein production. This was performed with inverse PCR followed by treatment with KLD Enzyme Mix (NEB).
Reporter system and series-parallel genetic architecture
The GFP reporter plasmid system developed by Rondon and Wilson44 was utilized for the single TF studies with a proximal operator and was also used as the series-parallel proximal architecture. This construction began with the pZS*22-sfGFP developed by Richards et al.40 (featuring a pSC101* origin, a kanamycin resistance marker, and lac O1+-regulated sfGFP reporter gene). The plasmid was linearized, excluding the promoter and operator elements, and a small fragment containing the constitutive trc promoter element55 (a hybrid of trp and lac UV5 promoters), a spacer sequence, and the operator sequence, was constructed via oligos, was inserted via CPEC. Each individual operator variant reporter was constructed via site-directed mutagenesis, beginning with the original proximal reporter plasmid (with the O1 operator). The wild-type rbsDACBK operator, identified by Shimada et al.56, was cloned into the engineered O1-regulated proximal reporter plasmid via site-directed mutagenesis. The core operator reporter system was adapted from the proximal. This was utilized for single TF studies with a core operator and was also used as the series-parallel core architecture. Using inverse PCR, the segment including the spacer sequence and operator was deleted, then site-directed mutagenesis was used to insert 17-bp of an operator sequence in the region between the −35 (TTGACA) and −10 (TATAAT) boxes of the trc promoter (similar to the original pLLac-O1 architecture). As with the proximal operator system, each individual operator variant reporter was constructed via site-directed mutagenesis, beginning with the original (which contained an O1 operator in the core position). The wild-type fruBKA operator, identified by Ramseier et al.57, was cloned into the engineered O1-regulated core reporter plasmid via site-directed mutagenesis. Attempts to assay with the fruBKA operator in the proximal position, as done with the rbsDACBK operator, did not show any repressibility or inducibility in assays with cells co-transformed with wild-type FruR.
Transcription factor vectors
To construct plasmids with multiple TF ORFs for use in genetic logic gates, additional plasmids were developed, as introduced by Rondon et al.25. All plasmids featured orthogonal origins of replications and selection markers. The first utilized the pLacI architecture. The pLacI plasmid was first linearized, visualized, and gel extracted (Qiagen), along with a fragment containing an amplified, second TF ORF, including a second constitutive LacI promoter. The plasmid was then assembled using the NEBuilder HiFi Kit (NEB). The constructs utilizing this architecture included the RA(1)/IA(5) and RA(1)/IA(9) dual-TF plasmids. A second TF vehicle plasmid was developed from the pSO vector generated in Rondon and Wilson44. This plasmid contains the PBR322 origin, an ampicillin (or carbenicillin) resistance marker, and a TF ORF driven by the constitutive LacI promoter. The original pSO vector was constructed by amplifying the AmpR coding region from pLS1 (Addgene #31490), then visualizing the gel extracting the fragment. This was then combined with the LacI (I+) ORF, amplified from pLacI, via splicing by overlap extension (SOE). Finally, the PBR322 origin was PCR amplified from the pET-28b vector (a gift from the Kane Lab), visualized, and gel extracted, before being combined with the SOE fragment via CPEC. To insert a second TF ORF into pSO, the vector was first linearized and the coding region for a second TF ORF was amplified. Both fragments were gel extracted, then the plasmid was assembled using the NEBuilder HiFi Kit (NEB). The constructs using this architecture included the R+/I+ dual-TF plasmid, including the variant with I+ under the LacIq promoter, and the FA(2)/IA(5) dual-TF plasmid. Modification of the promoter was performed on pSO prior to the insertion of a second TF, so as not to mutate the identical LacI promoter elements.
Two additional types of genetic architectures—series and parallel—were constructed for genetic logic gates. Briefly, series architectures consist of two sequential operators, one in the core position and one in the proximal, as defined by Cox et al.47, controlling one gene. Parallel architectures consist of two operators, in either the core or proximal positions, controlling two distinct ORFs. Both were adapted from Rondon et al.25, using the master architecture as parallel, with an orthogonal operator sequence in the unused position. For the series architecture, the starting point was the pZS*22-sfGFP plasmid. The plasmid, excluding the promoter and operator, was PCR amplified in two fragments, which were then visualized and gel extracted (Qiagen). The region upstream of the sfGFP gene containing the promoter element, operators, and the synthetic insulator RiboJ58 was synthesized via oligos. The trc promoter was used as a scaffold, but the region between the −35 and −10 boxes were replaced with 17-bp of an operator sequence, similar to the core architecture, above. The second operator was introduced in the proximal position 15-bp downstream of the end of the −10 box. This synthesized region was combined with the vector fragments by sequential SOE and CPEC reactions, resulting in a complete plasmid.
For the parallel architecture, the starting point was the series plasmid, described above. First, the plasmid was linearized, and this fragment was visualized and gel extracted. Next, the region upstream of the second copy of sfGFP was synthesized via oligos. In this case, we used the strong pL promoter element as the scaffold and replaced the region between the −35 (TTGACA) and −10 (GATACT) with 17-bp of an operator sequence. Again, the second operator was introduced in the proximal position, 15-bp downstream of the end of the −10 box. Another genetic insulator part was employed (this time, RiboJ1018, to introduce sequence diversity). The second copy of sfGFP (along with the strong terminator rrnb1) was amplified from the original pZS*22sfGFP, which was visualized and gel extracted, before being combined with the synthesized region via SOE. Finally, the linearized series architecture plasmid was combined with this newly synthesized fragment using the NEBuilder HiFi Kit. In both architectures, self-cleaving ribozyme genetic insulators were used to eliminate transfer function variability caused by differences in operator and 5′ UTR sequence25. When operator sequences needed to be altered to construct different logical operations, this was done using site-directed mutagenesis using inverse PCR followed by KLD enzyme treatment (NEB). The −10 box for the pL promoter element (GATACT) was mutated to the sequence GACTAT (from iGEM J23116) in the XNORp gate to account for difference in the OFFstates of repressors and anti-repressors. This was also done using inverse PCR, followed by KLD treatment.
Pairwise alignment was performed using the EMBL-EBI EMBOSS Needle online tool, with protein sequences from the UniProt database. Default settings were selected, with the matrix EBLOSUM62 and the gap and extend penalties set to 10.0 and 0.5, respectively. Multiple sequence alignment was similarly performed using the EMBL-EBI Clustal Omega web server, with protein sequences from the UniProt database (or, in the case of RbsR-L, Addgene #60773, and FruR-L, Addgene #60768).
Protein library creation
All site-saturation mutagenesis was performed using a protocol using inverse PCR with NNS degenerate codons. Following treatment with KLD Enzyme Mix (NEB), product was transformed into NEB 5-alpha Competent Cells. All transformants (i.e., the library) from selection plates were streaked and cultured for plasmid isolation and the DNA was sequenced to verify site-saturation coverage at the desired site. For mutagenesis to validate the transfer of allosteric networks between scaffolds, site-directed mutagenesis via inverse PCR was performed with primers to incorporate a single residue at a site (following screening to determine a desired residue) or, for incorporation of multiple mutations across the RCD, the pLacI vector and the fragment containing the mutant RCD were amplified, then combined via CPEC reaction.
The directed evolution library was constructed using a protocol adapted from Meyers et al.59. In one PCR, the pFruR-L vector was linearized excluding the region corresponding to the RCD (residue 63 through 334). While there is potential for mutation in the DBD or hinge region to contribute to allosteric inversion, we chose to constrain mutagenesis to the allosteric core, as done previously, in successful studies40,59. In another PCR, the FruR-L RCD region was subjected to error-prone PCR. A master library with 5- to 7-bp (3–5 residue) mutations, on average, over the 815-bp region (~0.7% error frequency), was constructed in a reaction with 1.25 U Taq DNA Polymerase (NEB), 1X Taq Mg-free buffer (NEB), 1.8 mM MgCl2 (NEB), 200 μM MnCl2 (Millipore Sigma), 0.4 μM dCTP (NEB), 0.4 μM dTTP (NEB), 0.08 μM dGTP (NEB), 0.08 μM dATP (NEB), 500 μM, each, forward and reverse primers, and 10 ng (4.2 fmol) of pFruR-L DNA, as template. The reaction was subjected to 95 °C for 3 min and 20 cycles of 95 °C for 30 s, followed by 68 °C for 5 min, and a final extension at 68 °C for 10 min. The vector and error-prone RCD insert were both visualized on gels, then extracted (Qiagen). The two fragments were combined via CPEC, then transformed into electrocompetent 3.32 cells. The library size was estimated to be on the order of 107 colony forming units (cfu). As above, all transformants were streaked from plates and plasmids were isolated. Libraries for screening were developed by performing site-directed mutagenesis on the master library to introduce the FS super-repressor mutations (namely, L95K, L95I, I96R, I96P).
Transformation into reporter cell line
All screening experiments were done in the 3.32 cell line. Isolated DNA constructs bearing TFs (on pSO or pLacI scaffolds) were co-transformed with isolated reporter plasmids and selected for on LB Agar plates. For site-saturation libraries, 96 colonies were picked and inoculated into LB with appropriate antibiotics and grown overnight – 96 were picked for chimera site-saturations, giving a 95% probability of all 32 codons (NNS; 4x4x2) were represented in the library.All transformants were cultured in LB (with antibiotics) overnight, at 37 °C shaking at 300 rpm, in preparation for screening. All screening for discovery was done using the O1 operator in the proximal position and the wild-type LacI DBD (YQR) on the TF.
The microplate assay protocol was taken from Richards et al.40. Briefly, colonies were inoculated in LB with relevant antibiotics (chloramphenicol for pLacI, kanamycin for the GFP reporter, carbenicillin for pSO) and grown overnight at 37 °C, shaking at 300 rpm. Cultures were then diluted 1:100 in supplemented M9 Minimal Media containing relevant antibiotics and ligands, as called for, in the wells of a 96-well, sterile, conical-bottomed microwell plate (Nunc). For single TF assaying, trials (variant + condition) were inoculated in replicates of 6, while for logic gates, replicates of 12 were used. Microplates were sealed with Breathe-Easier membranes (Midwest Scientific) to prevent evaporation. Microplates were grown at 37 °C, shaking at 300 rpm, for 20 h. Cultures were then transferred to the wells of a 96-well black-sided, clear bottomed assay plate (Costar). Fluorescence (characteristic to sfGFP—excitation: 485 nm, emission: 510 nm) and optical density (OD600) were measured by plate reader (Molecular Devices SpectraMax M2e). Data was collected with SoftMax Pro Software (Molecular Devices). Fluorescence intensity was normalized to cell density. Maximum fluorescence for all operator systems was measured by assaying a lacnull control, which is a variant in the pLacI vector with STOP codons at positions 2 and 3 of the LacI reading frame40. Hence, this plasmid produces no TF while exerting similar metabolic burden. This value was used for standardization across variants with the same operator. For logic gate assaying, all values were normalized to the highest value across all conditions. An averaged value for blank optical density and fluorescence was subtracted from all measured samples. Data analysis was performed in Microsoft Excel.
Conferring alternate DNA recognition
DBDs for each TF were mutated to ADR modules using inverse PCR followed by KLD enzyme treatment (NEB). Codons corresponding to positions Y17, Q18, and R22 on LacI were mutated to the respective 6 other DBD modules.
Cell cytometry analysis and sorting
Cultures were prepared similarly to those prepared for the microplate assay, and as done by Richards et al.40. Briefly, colonies were inoculated in LB with relevant antibiotics (chloramphenicol for pLacI, kanamycin for the GFP reporter, carbenicillin for pSO) and grown overnight at 37 °C, shaking at 300 rpm. Cultures were then diluted 1:100 in supplemented M9 Minimal Media containing relevant antibiotics and ligands, as called for, in sterile culture tubes (Nunc) and grown at 37 °C, shaking at 300 rpm, for 20 h. Each culture was then aliquoted such that the optical density (OD600) would be approximately equal to 0.2 (WPA Biowave CO8000 Cell Density Meter) in the final solution volume (typically 1 mL). Cells were then pelleted at 17,500 g for 2 min (Beckman Coulter Microfuge 18). The supernatant was then discarded, and cells were resuspended in PBS supplemented with 25 mM HEPES (Fisher Scientific), 1 mM EDTA (Millipore Sigma), and 0.01% (v/v) Tween20 (VWR Life Sciences), and again pelleted at 17,500 × g for 3 min. This wash step was repeated once before cells were finally resuspended in PBS supplemented with 25 mM HEPES and 1 mM EDTA.
Cytometry experiments were performed using a BD FACSAria Fusion flow cytometer (BD Biosciences) equipped with a 100 mW 488 nm laser for excitation, a 510/30 nm bandpass emission filter, and an 85 µm nozzle and data was collected using BD FACSDiva (BD Biosciences) software. Cells were interrogated measuring FITC-Area at flow rates between 10 and 30 and µL min−1. Events were gated on forward- and side-scatter and a threshold were set by side-scatter (5000), with doublets discriminated against using standard FSC-Area vs. -Height and SSC-Area vs. -Height plots. At least 25,000 events were recorded for cytometry analysis, while directed evolution libraries screened 20,000–80,000 events per sort. For directed evolution screening, cells were sorted directly into LB (with antibiotic), then cultured overnight at 37 °C, shaking at 300 rpm, in culture tubes (Nunc). This culture was then used to inoculate for another day of sorting after being prepared, as above. Sorts for FA phenotype consisted of two sequential sort steps: the first with the ligand, collecting the low-fluorescence mutants (FS and FA, screening out residual wild-type F+ and non-functional F−), the second without ligand, collecting the high-fluorescence mutants (discriminating between FS and FA). Gates for low and high fluorescence were determined from controls with known cytometry performance—i.e., lacnull, parent FS, and IA and IS variants from Richards et al.40. For the settings utilized, the high-fluorescence mutants were binned ≥104 FITC-A, while low-fluorescence mutants binned <104 FITC-A. Each individual sort step was repeated once to enrich the desired population and filter out false positives. Data was analyzed using the BD FlowJo (BD Biosciences) software package. See Supplementary Fig. 15 for gating and sorting strategies. Following the final sorts, cells were grown overnight at 37 °C, shaking at 300 rpm in LB with antibiotics, then plated on selection plates. Once grown overnight at 37 °C, individual colonies were screened via the microplate assay to verify phenotype and determine performance characteristics.
Phenotypes were evaluated by performing a two-tailed Student’s t-test, allowing for unequal variances, between the induced and uninduced states. The significance level was set to α = 0.001. Variants that saw no statistically-significant difference between the induced and uninduced states were XS and X− variants. Classification depended upon the magnitude of fluorescence output. If the fluorescence intensity was greater than 50% of the maximum for that operator variant, it was classified as X−; if less than 50% of the maximum, it was classified as an XS. For logic gates, a one-way ANOVA followed by a post-hoc Tukey HSD test was performed on the fluorescence outputs under the different ligand treatments to determine the statistical difference between groups. The significance level for logic gates was set to α = 0.01. For both phenotype evaluation and logic gates, Cohen’s d tests were performed to determine effects of sample size. For all cases, plots represent averages of biological replicates, with error bars representing ±1 standard deviation. See Source data File for p-values, effect size test parameters, and Tukey HSD test results.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.