Chemicals and reagents

All chemicals were of analytical grade and were purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.) unless otherwise indicated. Bacterial growth media were purchased from Thermo Fisher Scientific (Waltham, MA). Bacterial strains and all pure enzymes and related buffers were purchased from New England Biolabs (NEB, Ipswich, MA) unless otherwise indicated. KlenTaq1 was purchased from DNA Polymerase Technologies (St. Louis, MO). All oligonucleotides and gene blocks were obtained from Integrated DNA Technologies (IDT, Coralville, IA, U.S.A.). Oligonucleotide and gene block sequences are summarized in S1 Table.

Plasmids and cloning

PCR amplification of sequences for subsequent cloning was performed using Phusion DNA polymerase. Standard Gibson assembly techniques were used for all cloning unless otherwise noted. Coding sequences for shuffle-optimized KlenTaq DNA polymerase [25], Bst LF DNA polymerase [25], Taq DNA ligase (UniProtKB—B7A6G7), T5 Exonuclease (UniProtKB—P06229), MMLV reverse transcriptase (UniProtKB—P03355), Taq DNA polymerase [26], Phusion DNA polymerase [20, 21] and RTX thermostable reverse transcriptase [22] were cloned into pATetO 6xHis plasmid. This is an in-house designed plasmid based on the pASK-IBA37plus vector (IBA GmbH) from which the multiple cloning site, and Rop gene have been removed to improve plasmid copy number [25]. The plasmid also features a modified pAtetO promoter with a single point mutation to make it unidirectional. All enzyme coding sequences introduced into this vector were placed immediately downstream of the Factor X cleavage site. For some experiments, the coding sequences for wildtype and exonuclease deficient versions of RTX were cloned downstream of the T7 promoter in the pET21 vector (Sigma-Aldrich) [22]. Assembled plasmids were transformed into chemically competent Top10 E. coli and verified by Sanger sequencing at the Institute of Cellular and Molecular Biology Core DNA Sequencing Facility.

Production of lyophilized cellular reagents

Top10, BL21 and BL21 DE3 strains of E. coli were used to prepare lyophilized cellular reagents. Chemically competent BL21 and BL21 DE3 bacteria were freshly transformed with pATetO and pET21 constructs, respectively, prior to each instance of cellular reagent preparation. Top10 strains transformed with pATetO constructs and stored as glycerol stocks at -80°C were used to inoculate fresh cultures for cellular reagent preparation. Overnight 3 ml cultures of transformed bacterial strains were grown in 2X YT broth containing 100 μg/ml ampicillin. Subsequently, 50 ml sub-cultures at 1:200 dilution, unless otherwise specified, were initiated in Superior Broth^TM (Athena Environmental Sciences, Inc., Baltimore, MD, USA) containing 100 μg/ml ampicillin. Sub-cultures were incubated in 250 ml conical flasks at 37°C and constant 225 rpm agitation. Bacterial growth was monitored by measuring absorbance of 600 nm wavelength light.

Protein production was initiated by inducing transcription from the pATetO and the pT7 promoters by adding 200ng/ml anhydrotetracycline (aTC) or 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) to logarithm phase (typical A₆₀₀ = 0.4 to 0.7) cultures. The pATetO promoter was induced for 3h at 37°C, unless otherwise indicated. The pT7 promoter was induced for 18 h at 18°C.

After induction, bacteria were collected by centrifugation followed by washing once in cold 1X PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na₂HPO₄, 1.47 mM KH₂PO₄, pH 7.4). The bacterial pellets were resuspended in cold 1X PBS at a density of A₆₀₀ = 3.5 to 6.5. Some 2x10⁸ aTC-induced bacteria and 2x10⁷ IPTG-induced bacteria (estimated from the A₆₀₀ value using the relation 0.5 optical density = 5x10⁸ bacteria/ml) were aliquoted into individual 0.2 ml PCR tubes and frozen at -80°C overnight prior to lyophilization for 3 h at 197 mTorr and -108°C using the automated settings in a VirTis Benchtop Pro lyophilizer (SP Scientific, Warminster, PA, USA). Lyophilized cellular reagents were stored with desiccant at room temperature, 37°C, or 42°C until use.

Purification of RTX reverse transcriptase

RTX Exo- polymerase was expressed and purified in house following the protocol of Ellefson et al (2016) [22]. Briefly, BL21 DE3 bacteria harboring the pET21-RTX Exo- polymerase containing plasmid was grown overnight in Superior Broth^TM at 37°C. Cells were then diluted 1:200, and protein production was induced with 1 mM IPTG during mid-log phase at 18°C for 16–18 hrs. Harvested cells were flash-frozen and lysed by sonication in 10 mM phosphate, 100 mM NaCl, 0.1 mM EDTA, 1 mM DTT, 10% glycerol, pH 7 buffer containing protease inhibitor (Sigma-Aldrich). Cell lysate was then centrifuged at 40,000g for 45 min at 4°C. Cleared cell lysates were heated at 85°C for 25 min, cooled on ice for 20 min, and spun again at 20,000g for 15 min. Supernatant obtained after centrifugation was filtered using 0.2 μm filters. The filtrate was then passed over an equilibrated heparin column (GE Life Sciences, Pittsburgh, PA, USA), and eluted along a sodium chloride gradient. Polymerase fractions were collected and dialyzed into Buffer A [22]. Enzymes were further purified using an SP column (GE Life Sciences) and again eluted along a salt gradient. Pooled fractions were then applied to a Sephadex 16/60 size exclusion column (GE Life Sciences), concentrated, and dialyzed into storage buffer (50 mM Tris-HCl, 50 mM KCl, 0.1 mM EDTA, 1 mM DTT, 0.1% Nonidet P40, 0.1% Tween-20, 50% glycerol, pH 8.0). Purified RTX Exo- polymerase was quantified by Pierce BCA protein assay kit (Thermo Fisher Scientific).

Overlap extension assay using Taq DNA polymerase

Taq DNA polymerase-expressing Top10 E. coli cells were cultured and processed as cellular reagents as described above. Prior to freeze drying, an aliquot containing 2 x 10⁸ of these freshly cultured cells was resuspended in 30 μL 1X PBS and centrifuged for 1 min at 13,000 rpm. The resulting supernatant was collected in a fresh tube while the cell pellet was resuspended in 30 μL water. Taq DNA polymerase activities were measured in 3 μL aliquots of the supernatant and in 2 x 10⁷ prepared cells (contained in 3 μL aliquots) using overlap extension assays executed as follows. Forty seven microliter reactions containing 2 μM each of overlapping oligonucleotides OE.FWD and OE.REV, 1X Thermopol buffer (NEB) (20 mM Tris-HCl, 10 mM (NH4)₂SO₄, 10 mM KCl, 2 mM MgSO₄, 0.1% Triton®-X-100, pH 8.8), and 0.2 mM deoxyribonucleotides (dNTPs) were assembled and heated to 95°C for 1 min followed by cooling on ice for 2 min. Then, 3 μL supernatant or 3 μL cells (2x10⁷) were added to the reactions, which were then incubated for 1 h at 37°C, 42°C, 65°C, or 75°C. Negative control reactions were performed in the same manner with the exception that water was used instead of OE.FWD and OE.REV oligonucleotide templates. A second aliquot of 2 x 10⁸ cells was first frozen in 1X PBS at– 80°C prior to testing for Taq DNA polymerase activity in cells and supernatant as described above. A third aliquot of 2 x 10⁸ cells was frozen in 1X PBS and then lyophilized prior to rehydration with 30 μL water and testing for Taq DNA polymerase activity in oligonucleotide extension assays. Supernatant could not be separated from this rehydrated sample under the centrifugation conditions described above. All oligonucleotide extension products were analyzed by ethidium bromide agarose gel electrophoresis.

Endpoint PCR using fresh broth culture of cellular reagents

BL21 DE3 bacteria transformed with RTX Exo- polymerase expression plasmid were grown to logarithm phase in Superior Broth^TM and induced with 1 mM IPTG as described in Section 2.3. One milliliter culture of induced cells was centrifuged at 16,000g for 1 min. Supernatant was removed and bacteria were resuspended in 1 ml PBS. 1 μl of this neat or 1:10 diluted bacterial suspension was added to a 20 μl PCR reaction containing 10 ng (1 x 10⁸ copies) of Chlamydia trachomatis 16s rDNA templates in 1X PCR proof reading assay buffer (60 mM Tris-HCl (pH8.4), 25 mM (NH4)₂SO₄, 10 mM KCl and 1 mM MgSO₄) supplemented with 0.5 mM dNTPs and 400 nM each of forward (CT.F) and reverse (CT.R) primers. In positive control PCR reactions, RTX Exo- expressing bacterial suspension was replaced with 1 μl of purified RTX Exo- polymerase (0.2 mg/ml stock) or with commercially available KOD DNA polymerase (Sigma Aldrich, St. Louis, MO). Reactions were incubated at 95°C for 5 min followed by 25 cycles of 20 sec at 95°C, 20 sec at 55°C, and 20 sec at 68°C. PCR products were analyzed by electrophoresis through a 1.5% agarose gel prepared in 1X TAE Buffer (22 mM Tris, 180 mM-borate, 5 mM EDTA pH 8.3). Some 0.5 μg/ml of ethidium bromide was included in the gel to visualize the DNA bands under UV light. After running the gel at 80 Volts for 30 min, bands were visualized using ChemiDoc Imager (Bio-Rad).

Endpoint PCR using Taq DNA polymerase lyophilized cellular reagents

Endpoint PCR reactions were assembled in 50 μL volumes containing zero or 10 ng of a pCR2.1-FluB plasmid template along with a final concentration of 500 nM each of pCR.FWD and pCR.REV primers. Amplification was performed in 1X Thermopol buffer (NEB) containing 0.2 mM dNTPs and 3 μL (2 x 10⁷ cells) of Taq DNA polymerase cellular reagent rehydrated in 30 μL water immediately prior to use. Following an initial 10 min incubation at 95°C, the reactions were subjected to 30 thermal cycles of 30 sec at 95°C, 30 sec at 55°C, and 1 min at 72°C. Ten microliters of the resultant PCR products were analyzed by agarose gel electrophoresis.

Real-time LAMP-OSD

LAMP-OSD reaction mixtures were prepared in 25 μl volume containing indicated amounts of human glyceraldehyde-3-phosphate dehydrogenase (gapd) DNA templates along with a final concentration of 1.6 μM each of BIP and FIP primers, 0.4 μM each of B3 and F3 primers, and 0.8 μM of the loop primer. Amplification was performed in 1X isothermal buffer (NEB) (20 mM Tris-HCl, 10 mM (NH4)₂SO₄, 10 mM KCl, 2 mM MgSO₄, 0.1% Triton X-100, pH 8.8) containing 0.8 M betaine, 0.8 mM dNTPs, 2 mM additional MgSO₄, 16 units of pure Bst 2.0 DNA polymerase, and 100 nM of OSD reporter. Reporters were prepared for use in LAMP assays by annealing 100 nM fluorophore-labeled OSD strands with a 5-fold excess of the quencher-labeled OSD strands by incubation at 95°C for 1 min followed by cooling at the rate of 0.1°C/sec to 25°C). In some LAMP-OSD assays commercial Bst 2.0 was substituted either with 16 units of commercial Bst-LF DNA polymerase or with 3 μl (2 x 10⁷ cells) of Bst-LF expressing lyophilized BL21 cellular reagents (rehydrated prior to use in 30 μl water). For real-time signal measurement these LAMP-OSD reactions were transferred into a 96-well PCR plate, which was incubated in a LightCycler 96 real-time PCR machine (Roche, NC, U.S.A.) maintained at 65°C for 90 min. Fluorescence signals were recorded every 3 min and analyzed using the LightCycler 96 software.

Quantitative PCR

KlenTaq DNA polymerase qPCR reactions were prepared in 25 μl volume containing indicated amounts of Chlamydia trachomatis 16S DNA templates along with a final concentration of 0.4 μM each of forward (CT.F) and reverse (CT.R) primers. Amplification was performed in 1X KlenTaq1 buffer (DNA Polymerase Technology) (50mM Tris-Cl pH 9.2, 16 mM ammonium sulfate, 0.05% Brij 58, and 3.5 mM magnesium chloride) containing 0.4 mM dNTPs, 0.2 μl of pure KlenTaq1 DNA polymerase, and 1X EvaGreen intercalating dye (Biotium, Freemont, CA). In some qPCR assays, commercial KlenTaq1 was substituted with 3 μl (2 x 10⁷ cells) of KlenTaq expressing lyophilized BL21 cellular reagents (rehydrated prior to use in 30 μl water). For real-time signal measurement these qPCR reactions were transferred into a LightCycler 96 real-time PCR machine and subjected to 10 min at 95°C followed by 45 cycles of 10 sec at 95°C (denaturation), 30 sec at 55°C (annealing) and 30 sec at 72°C (extension). Fluorescence signals were recorded during the extension step in each cycle. Following qPCR amplicon melting curve analysis was performed. All data were analyzed using the LightCycler 96 software.

Taq DNA polymerase TaqMan qPCR reactions were prepared in 25 μl volume containing indicated amounts of Zika virus NS5 DNA templates along with a final concentration of 0.32 μM each of forward (Zika-4481_F) and reverse (Zika-4552c) primers [27]. Amplification was performed in 1X Thermopol buffer (NEB) containing 0.4 mM dNTPs, 2.5 units of Taq DNA polymerase, and 80 nM TaqMan probe (Zika-4507c-FAM) [27]. In some assays commercial Taq DNA polymerase was substituted with 3 μl (2 x 10⁷ cells) of Taq DNA polymerase expressing lyophilized BL21 cellular reagents (rehydrated prior to use in 30 μL water). For real-time signal measurement these TaqMan qPCR reactions were transferred into a LightCycler 96 real-time PCR machine and subjected to 10 min at 95°C followed by 45 cycles of 15 sec at 95°C (denaturation) and 30 sec at 55°C (annealing and extension). Fluorescence signals were recorded during the annealing/extension step in each cycle. All data were analyzed using the LightCycler 96 software.

RTX reverse transcriptase qPCR reactions were prepared in 20 μl volume containing indicated amounts of Chlamydia trachomatis or Zika virus-derived DNA templates along with a final concentration of 200 nM each of forward and reverse primers (CT.F/R and Zika-255-F/Zika-256-R, respectively). Amplification was performed in 1X PCR proof reading assay buffer (60 mM Tris-HCl (pH8.4), 25 mM (NH4)₂SO₄, 10 mM KCl and 1 mM MgSO₄) containing 0.5 mM dNTPs, 1.5 M Betaine, 1X EvaGreen dye and 80 ng of RTX Exo- polymerase. In some assays, purified RTX polymerase was substituted with 5μl (2X10⁶ cells) of lyophilized cellular reagent (rehydrated prior to use in 50 μl water). For real-time signal measurement these qPCR reactions were transferred into a LightCycler 96 real-time PCR machine and subjected to 5 min at 95°C followed by 45 cycles of PCR. Chlamydia trachomatis template was cycled through 20 sec at 95°C, 20 sec at 55°C and 20 sec at 68°C while Zika template was cycled through 30 sec at 95°C, 30 sec at 55°C and 30 sec at 68°C. Fluorescence signals were recorded during the extension (68°C) step in each cycle. All data were analyzed using the LightCycler 96 software.

Two-step quantitative reverse transcription (RT) PCR

Indicated amounts of in vitro transcribed and polyacrylamide gel purified Zika virus NS5 RNA templates were mixed with 10 μM reverse (Zika-4552c) primers and 1 mM dNTP in a total volume of 10 μl. Primer template annealing was performed by incubating the solutions at 65°C for 5 min followed by 2 min on ice. Reverse transcription was initiated by adding a 10 μl solution containing 2X MMLV RT buffer (NEB) (100 mM Tris-HCl, 20 mM DTT, 150 mM KCl, 6 mM MgCl₂, pH 8.3), 8 units of RNase inhibitor and 3 μl (2 x 10⁷ cells) of MMLV reverse transcriptase expressing lyophilized cellular reagents (rehydrated prior to use in 30 μl water). Following 1 h of reverse transcription at 42°C, 5 μl of the resulting cDNA-containing solution was analyzed by TaqMan qPCR using Taq DNA polymerase expressing lyophilized BL21 cellular reagents as described above.

One-pot quantitative reverse transcription (RT) PCR using RTX polymerase

One-pot RT-qPCR reactions using pure or cellular RTX Exo- reagent and indicated copies of Zika virus-derived RNA templates were assembled using the same procedure as RTX qPCR described above but with the addition of 10 mM DTT. For real-time signal measurement these RT-qPCR reactions were transferred into a LightCycler 96 real-time PCR machine and subjected to 68°C for 30 min followed by 5 min at 95°C prior to 45 cycles of 30 sec at 95°C, 30 sec at 55°C and 30 sec at 68°C. Fluorescence signals were recorded during the extension (68°C) step in each cycle. All data were analyzed using the LightCycler 96 software.

Cellular PCR (cPCR) using lyophilized Phusion cellular reagents

Bacteria containing target DNA sequences were grown overnight at 37°C and 250 rpm in 3 ml 2X YT broth containing the appropriate antibiotics for selective pressure. Following overnight growth, cultures with A₆₀₀ of 5 to 6 were diluted 1:10 in sterile water. Two microliters of this diluted culture was added to cPCR reaction to initiate amplification. cPCR reactions were assembled in a total volume of 50 μl containing 1X HF buffer (NEB), 0.2 mM dNTP mix, 0.2 μM each of forward and reverse primers and 3 μl of Phusion cellular reagent (prepared by rehydrating 2x10⁸ lyophilized Top10 E. coli expressing Phusion DNA polymerase in 30 μl water). cPCR reactions were incubated at 95°C for 10 min followed by 30 cycles of 30 sec at 95°C, 30 sec at 60°C, and 5 min at 72°C. cPCR products were analyzed by agarose gel electrophoresis and used directly for Gibson assembly. A portion of cPCR products were subjected to agarose gel purification using Wizard SV gel purification kit (Promega, Madison, WI, USA) prior to use in Gibson assembly.

Gibson assembly and transformation of chemically competent bacteria

Twenty microliters Gibson assembly reactions were assembled by mixing vectors and inserts in 1X Gibson assembly buffer (0.1 M Tris-HCl, pH 7.5, 0.01 M MgCl₂, 0.2 mM dGTP, 0.2 mM dATP, 0.2 mM dTTP, 0.2 mM dCTP, 0.01 M DTT, 5% (w/v) PEG-8000, 1 mM NAD) supplemented with pure enzymes or with cellular reagents. Assemblies using pure enzymes contained 0.08 units of T5 exonuclease (NEB), 0.5 units of Phusion DNA polymerase (NEB) and 80 units of Taq DNA ligase (NEB). For Gibson assemblies using cellular reagents the pure enzymes were substituted with individual Top10 E. coli cellular reagents expressing Taq DNA Ligase, Taq DNA polymerase, and T5 exonuclease.

Lyophilized cellular reagents were prepared as follows for use in Gibson assemblies. 2x10⁸ Top10 lyophilized cellular reagents expressing Taq DNA polymerase, T5 exonuclease, or Taq DNA ligase were rehydrated using 30 μl of water. The rehydrated T5 exonuclease cellular reagent was diluted 1:100 in water followed by addition of 1.5 μl aliquot per Gibson reaction. Rehydrated Taq DNA polymerase and Taq DNA Ligase cellular reagents were incubated at 75°C for 10 min. Three microliters of the heat-treated cellular Taq DNA Ligase was directly added to cellular Gibson assemblies. Heat-treated Taq DNA polymerase cellular reagents were diluted 1:10 in water prior to addition of 1.5 μl aliquot per Gibson assembly.

Linearized vectors and inserts for Gibson assemblies were mixed in the following ratios: Two part assemblies of double stranded gene block (gBlock) inserts (IDT) and linearized vector (agarose gel purified PCR product) contained equimolar ratio of vector (30 ng) and gBlock (3 ng). Two part assemblies of cPCR-amplified vectors and inserts contained 2.5 μl each of unpurified or gel-purified vector and insert cPCR products. Three-part cPCR vector and insert assemblies contained 1.5 μl of cPCR vector fragment 1, 1.5 μl of cPCR vector fragment 2 and 2 μl of insert cPCR.

Negative controls for Gibson assemblies included vectors and inserts in 1X Gibson buffer without any pure enzymes or cellular reagents. All Gibson assemblies were incubated at 50°C for 1 h prior to transformation.

One microliter of each Gibson assembly was transformed directly into chemically-competent Top10 E. coli bacteria. Briefly, 50 μl home-made competent bacteria [28] were mixed with the Gibson assembly and incubated on ice for 15 min. Following a 30 sec heat-shock at 42°C and a 2 min incubation on ice bacteria were allowed to recover for 1 h in 250 μl SOC medium (0.5% Yeast Extract, 2% Tryptone, 10 mM NaCl, 2.5 mM KCl, 10 mM MgCl₂, 10 mM MgSO₄, 20 mM Glucose) at 37°C on a rotator. All bacteria were collected by centrifugation and plated on 2X YT agar plates containing appropriate antibiotic selection. Following 18 h at 37°C bacterial colonies were counted to determine efficiency of Gibson assembly. Representative colonies were verified by Sanger sequencing.

Statistical analysis

A minimum of three biological replicates with duplicate or triplicate embedded technical replicates were performed. ANOVA testing was used to determine whether cellular reagents significantly affect the Cq value of a reaction. Two different ANOVA tests were performed. One for PCR like reactions and one for LAMP based reactions. For LAMP based reactions, cellular Bst-LF was compared to purified Bst-2.0. Effect sizes were determined using linear regression. The Cq response was modeled as a function of template copy number, the reaction type (e.g. RTX-qPCR, EVG-qPCR KlenTaq, etc.), and reagent type (purified or cellular). The effect size of cellular reagents on Cq value was determined using the “effects” package from the Comprehensive R Archive Network (CRAN) repository. Reactions with either zero template copies or no observed Cq values were omitted from analysis. All models were built using RStudio version 1.0.136. The data modeled, scripts, and a more detailed summary of the statistical analysis are included as supplements (S1 Statistical Supplement).

Development of cellular reagents for PCR

To determine the feasibility of using bacteria that express heterologous proteins directly as reagents in molecular biology reactions we carried out PCR using fresh cultures of bacteria expressing RTX Exo- polymerase. We observed bands of the expected size (~ 200 bp), similar to a positive control that contained purified RTX Exo- polymerase or KOD DNA polymerase (S1 Fig). These results suggest that bacteria expressing a polymerase enzyme could indeed be directly used as reagents without requiring prior purification of the polymerase.

While the use of fresh cultures for promulgating reactions presents many opportunities, it can also prove cumbersome. Therefore, we attempted to show that cellular reagents could also be stored and used. Overexpressing bacteria were either stored at -80°C or lyophilized, prior to functional comparison with pure enzymes in qPCR reactions. Enzyme activities of cellular reagents versus corresponding pure enzymes were compared by measuring the Cq values (time to detection) for the same number of template copies. Since there are typically yield losses during protein purification, it is difficult to accurately measure the physical amount of enzyme present in cellular reagents. Therefore, we compared the enzyme activity of defined numbers of RTX Exo- bacterial cellular reagents with the activity of corresponding pure RTX enzymes used at an experimentally determined optima of 80 ng / reaction. Similar C_q values for target detection were obtained irrespective of whether 2 x 10⁶ lyophilized cells/reaction or 80 ng of purified polymerases were used (S2 Fig). Frozen cellular reagents or 2 x 10⁵ lyophilized cells/reaction yielded somewhat higher (less sensitive) C_q values (S2 Fig).

Given that conventional wisdom would suggest that the addition of a highly complex mixture of enzymes and metabolites (i.e., a cell, lyophilized or lysed) to a reaction might have resulted in significant background amplification, an examination of additional negative controls was especially important. While purified enzyme did not generate spurious amplicons, frozen or lyophilized bacteria expressing RTX Exo- polymerase generated a small amount of non-specific signal in the absence of templates. These non-specific signals, however, could be readily distinguished from specific amplicons by performing melt curve analysis. Lyophilized BL21 DE3 bacteria that do not harbor the RTX Exo- polymerase generated linear background curves without measurable Cq values when presented with 1 x 10⁸ templates. Since lyophilized cellular reagents demonstrated comparable qPCR performance as pure enzymes and were stable at room temperature storage for at least 80 days (S3 Fig) we chose to focus on the development of lyophilized cellular reagent toolkits for potential diagnostic and molecular biology applications. Lyophilized cellular reagents have the added advantage of biosafety; E. coli lyophilized without excipients (such as trehalose) do not retain viability (S4 Fig) [29, 30], and therefore are safer for distribution and use.

It was possible that cellular reagents might be limited to the RTX enzyme, which had been evolved via emulsion-based selections that involved whole bacteria, and was therefore pre-optimized to function in the presence of bacterial components. Moreover, it was unclear whether cellular reagents required bacterial lysis, such as by heat treatment [31] due to cycling through 95°C during PCR, to facilitate enzyme access, a step that might limit the development of mesophilic cellular reagents.

Therefore, we attempted to determine if other enzymes commonly used for PCR could be similarly repackaged as simpler ready-to-use cellular reagents. E. coli cells overexpressing Taq DNA polymerase were washed in PBS and assessed for enzyme activity in three different conditions: fresh cells, cells frozen at -80°C, or lyophilized cells. Cells were tested isothermally by single step overlap extension assays at four different temperatures– 37°C, 42°C, 65°C, and 75°C–to evaluate accessibility. The PBS supernatants leftover after pelleting fresh or frozen cells were also tested for polymerase activity. Rehydrated lyophilized bacteria could not be separated from supernatant under similar centrifugation conditions. Our results demonstrate that most of the Taq DNA polymerase activity is associated with bacterial cells (S5 Fig). Small amounts of activity evident in the supernatants is likely due to contaminating cells and/or cellular components that were not removed by centrifugation. Amplicon output of fresh and frozen cells was similar to that of pure Taq DNA polymerase only at temperatures ≥65°C, suggesting that these cells likely have intact cell walls that restrict enzyme accessibility and hence observable activity at more mesophilic temperatures. In contrast, the yield of amplicons generated by rehydrated lyophilized cellular reagents was similar to pure enzyme at all temperatures, suggesting unhindered access to the enzyme payload of cellular reagents. It seems osmotic shock may have the same effect as heat treatment for lyophilized cells.

Microscopic examination of Gram-stained rehydrated cellular reagents revealed that most E. coli cells lyophilized in 1X PBS were not dispersed but appeared hollow, unlike similarly stained fresh cells or cells lyophilized in water (S6 Fig). In contrast, upon heating for 5 min at 95°C rehydrated cells that had been lyophilized in 1X PBS were found to disintegrate while, most rehydrated cells that had been lyophilized in water retained their shape even after heat treatment. These observations suggests that E. coli freeze-dried in 1X PBS might have porous walls that allow intermingling of cellular and external reaction contents upon rehydration.

Enzymes other than RTX reverse transcriptase should be compatible with a cellular reagent format, especially since bacterial lyophilization in the presence of PBS is sufficient for enzyme access and additional bacterial lysis procedures are not required. In support of this hypothesis, Taq DNA polymerase cellular reagents also do not require cold storage; at the time of this publication these reagents remained functional even after 3 weeks at temperatures as high as 42°C (S7 Fig). Indeed, Taq DNA polymerase cellular reagents could replace pure commercial enzymes in the TaqMan qPCR assay, one of the most commonly used gold standards for molecular diagnostic procedures. A previously described molecular diagnostic assay for Zika virus detection [27] was used as an exemplar, albeit with synthetic DNA, instead of RNA, templates. Defined numbers of Taq DNA polymerase cellular reagents or standard amounts (2.5 units according to the manufacturer) of commercial Taq DNA polymerase were used to amplify the same number of template copies. Polymerase activities were compared by measuring the respective Cq values for the same number of templates. Some 2 x 10⁷ bacteria bearing Taq DNA polymerase demonstrated similar amplification efficiencies, detection limits, C_q values (time to detection), signal amplitudes, and absence of non-specific signals as added, 2.5 units of purified commercial Taq DNA polymerase (Fig 1). Most importantly, nucleic acid degradation was not evident in assays carried out with cellular reagents, likely due to inhibition and/or degradation of endogenous nucleases upon thermal cycling. Just as the preparation of thermostable enzymes in mesophilic cells is often abetted by the denaturation or aggregation of non-thermostable proteins [32], the same phenomenon can assist with molecular diagnostics.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. TaqMan qPCR analysis using lyophilized Taq DNA polymerase cellular reagents.

Indicated copies of synthetic DNA templates derived from Zika virus genomic sequence were amplified using 2.5 units of pure commercial Taq DNA polymerase (panels a and b) or 2 x 10⁷ cells of rehydrated cellular reagents expressing Taq DNA polymerase (panels c and d). Amplification was assessed in real-time by measuring increase in TaqMan probe fluorescence over time. Representative amplification curves generated using the “Abs quant” analysis in the LightCycler 96 software are depicted in panels a and c. Amplification curve colors distinguish starting template copies–yellow: 60,000 template copies; green: 6000 template copies; blue: 600 template copies; red: 60 template copies; and gray: no template control. These curves depict the real-time kinetics of PCR amplification mediated by pure versus cellular reagents. The corresponding standard curve analyses performed using the “Abs quant” protocol in the LightCycler 96 software are depicted in panels b and d, respectively. Standard curve analyses data for comparing amplification efficiency, linearity, and error are tabulated as insets.

https://doi.org/10.1371/journal.pone.0201681.g001

Single enzyme cellular reagents clearly were able to be used as molecular diagnostics, raising the possibility that multiple enzymes that worked in parallel could be delivered via cells. We attempted to set up a TaqMan qPCR assay starting with RNA templates and cellular reagents. RNA templates were first reverse transcribed into complementary DNA (cDNA) using MMLV reverse transcriptase-expressing lyophilized bacteria, and then without further purification TaqMan qPCR was carried out using Taq DNA polymerase cellular reagents. The two-step cellular reagent protocol could unambiguously detect templates bearing portions of the Zika virus with C_q values that correlated closely with the expected 10-fold differences between various samples (Fig 2). RNA samples that were not subjected to reverse transcription also yielded detectable C_q values but these were 6.5 to 7.5 units higher than the C_q values obtained with the cDNA samples, as was expected in the absence of any explicit DNase treatment. Non-specific signal in the absence of templates was not observed.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. RNA detection by two-step reverse transcription TaqMan qPCR using cellular reagents for MMLV RT and Taq DNA polymerase.

Indicated copies of synthetic RNA template derived from Zika virus genomic sequence were tested using 2 x 10⁷ cells each of MMLV RT and Taq DNA polymerase lyophilized cellular reagents. Amplification was assessed in real-time by measuring increase in TaqMan probe fluorescence over time. Representative amplification curves generated using the “Abs quant” analysis in the LightCycler 96 software are presented. Color of the traces indicate presence (black traces) or absence (red traces) of MMLV RT cellular reagents, or the absence of templates (blue traces). The corresponding derivation of template copies from Cq analyses are tabulated. Cq values were converted to template copies using standard curve analyses of the same RNA samples with commercial qRT-PCR master mix (S8 Fig).

https://doi.org/10.1371/journal.pone.0201681.g002

Finally, since many diagnostic protocols use DNA intercalating fluorophores for measuring amplicon accumulation we developed EvaGreen qPCR mixes containing lyophilized BL21 cellular reagents expressing KlenTaq or RTX Exo- DNA polymerase. In standard curve analyses of Chlamydia trachomatis DNA templates, 2 x 10⁷ cells of KlenTaq polymerase-expressing cellular reagents demonstrated similar amplification efficiencies, Cq values, and detection limits as the standard pure enzyme amount of 0.2 μL / reaction suggested by the manufacturer. Similarly, 2 x 10⁶ RTX Exo- cellular reagents demonstrated similar amplification efficiencies, C_q values, and detection limits with Zika virus-derived DNA templates, as optimal amount of 80 ng/ reaction of the pure enzyme (Figs 3 and 4). In fact, a one-step qRT-PCR mix built using cellular reagents expressing the RTX thermostable reverse transcriptase could accurately quantify Zika virus-derived RNA templates in a one-pot reaction (Fig 5). Despite the fact that a TaqMan probe was not involved, negligible non-specific signal was observed with either of these approaches. Moreover, target-derived amplicons could be readily identified by their characteristic melting temperatures. The slight increase in amplicon melting temperature observed in reactions containing cellular reagents might be due to an accompanying increase in ionic concentration and molecular crowding [33, 34].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. EvaGreen qPCR analysis using KlenTaq DNA polymerase expressing cellular reagents.

Indicated copies of synthetic Chlamydia trachomatis DNA template were amplified by PCR using 0.2 μL of pure commercial KlenTaq DNA polymerase (panels a, b, and c) or 2 x 10⁷ cells of KlenTaq cellular reagents (panels d, e, and f). Amplicon accumulation was assessed in real time by measuring increase in EvaGreen fluorescence. Panels a and d depict representative amplification curves generated using the “Abs quant” analysis in the LightCycler 96 software. Colors of the curve traces indicate starting numbers of template copies–black: 6x10⁶ template copies; blue: 6x10⁵ template copies; red; 6x10⁴ template copies; green: 6x10³ template copies; pink: 600 template copies; purple: 60 template copies; yellow: 6 template copies; and cyan: no templates. Taken together, these curves demonstrate the real-time kinetics of PCR amplification. Since EvaGreen is a non-specific DNA intercalating dye, the fidelity of amplicon generation was verified by determining their melting temperatures (panels b and e) using the “Tm calling” analysis protocol in the LightCycler 96 software. Color coding of the curves is the same as in panels a and d. The overlapping melting temperature peaks of amplicons generated from 6 x 10⁶ to 60 copies of templates are indicative of correctly amplified PCR products. Amplification curves observed in the presence of 0 to 6 template copies are non-specific as evident from their different melting temperatures peaks of these amplicons. Standard curve analyses performed using the “Abs quant” protocol in the LightCycler 96 software are depicted in panels c and f, respectively, and data for amplification efficiency, linearity, and error are tabulated as insets.

https://doi.org/10.1371/journal.pone.0201681.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. EvaGreen qPCR analysis using RTX Exo- DNA polymerase expressing cellular reagents.

Indicated copies of synthetic Zika virus derived DNA template were amplified by PCR using 80 ng of pure RTX Exo- DNA polymerase (panels a, b, and c) or 2 x 10⁶ cells of RTX Exo- cellular reagents (panels d, e, and f). Amplicon accumulation was assessed in real time by measuring increase in EvaGreen fluorescence. Representative amplification curves using 10⁸, 10⁷, 10⁶, 10⁵, 10⁴, 10³, 10², 10 and 0 template DNA copies are shown in panels a and d. ‘NTC’ refers to no template control. These curves were generated using the “Abs quant” analysis protocol in the LightCycler 96 software. The corresponding amplicon melting temperature analyses performed using the “Tm calling” protocol in the LightCycler 96 software are shown in panels b and e. The melting temperature peaks of target-derived amplicons are distinct from those of non-specific amplicons generated in the absence of templates. Standard curve analyses performed using the “Abs quant” protocol in the LightCycler 96 software are depicted in panels c and f. Standard curve analyses data for comparing amplification efficiency, linearity, and error are tabulated as insets.

https://doi.org/10.1371/journal.pone.0201681.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. EvaGreen qRT-PCR analysis using RTX Exo- DNA polymerase expressing cellular reagents.

Indicated copies of synthetic Zika virus derived RNA template were amplified by RT-PCR using 80 ng of pure RTX Exo- DNA polymerase (panels a, b, and c) or 2 x 10⁶ cells of RTX Exo- cellular reagents (panels d, e, and f). Amplicon accumulation was assessed in real time by measuring increase in EvaGreen fluorescence. Representative amplification curves using 10⁸, 10⁷, 10⁶, 10⁵, 10⁴, 10³, 10², 10 and 0 template RNA copies are shown in panels a and d. These curves were generated using the “Abs quant” analysis protocol in the LightCycler 96 software. ‘NTC’ refers to no template control. The corresponding amplicon melting temperature analyses performed using the “Tm calling” protocol in the LightCycler 96 software are shown in panels b and e. The melting temperature of non-specific amplicons generated in the absence of templates is distinct from target-derived amplicons. Standard curve analyses performed using the “Abs quant” protocol in the LightCycler 96 software are depicted in panels c and f. Standard curve analyses data for comparing amplification efficiency, linearity, and error are tabulated as insets.

https://doi.org/10.1371/journal.pone.0201681.g005

Overall, these results indicate that cellular reagents are suitable for a variety of PCR formats. Under the reported conditions substitution of commercial reagents with cellular reagents in qPCR resulted in a statistically significant but minute increase of Cq by 1.6 units (S1 Statistical Supplement). This small difference should not be a deterrent to the adoption of cellular reagents, especially since the same number of template copies could be detected with cellular or purified commercial reagents, with no interference from non-specific amplicons.

Isothermal nucleic acid amplification using cellular reagents

Beyond PCR, it should be possible to carry out other reactions using cellular reagents. To demonstrate the general utility of the concept, cellular reagents expressing Bst-LF, the Bacillus stearothermophilus strand displacing DNA polymerase (large fragment) commonly used for isothermal nucleic acid amplification reactions were assayed for their ability to carry out LAMP-OSD (S9 Fig). Similar to 16 units of pure commercial Bst 2.0 enzyme, 2 x 10⁷ Bst-LF cellular reagents could amplify as few as 60 copies of human glyceraldehyde-3-phosphate dehydrogenase gene target, within 60 min (Fig 6). Although, Bst 2.0 is an in silico designed polymerase engineered for greater amplification speed and yield than Bst-LF (NEB), 2 x 10⁷ cellular Bst-LF reagents were only ~2 Cq slower than 16 units of commercial Bst 2.0 (S1 Statistical Supplement). This level of performance by Bst-LF cellular reagents is especially impressive considering the fact that most users typically apply only 8 units of Bst 2.0 per reaction to reduce reaction cost [35]. However, use of 10-fold lower amounts of Bst-LF cellular reagents failed to generate any LAMP-OSD signals (data not shown). Nonetheless, as long as a sufficient number of bacteria were added, the production process for the creation of lyophilized cellular reagents was robust to perturbations such as bacterial sub-culture initiation density, optical density at logarithm phase induction of protein expression, alteration of expression platform, and induction duration (S10 Fig). Furthermore, similar to Taq DNA polymerase, Bst DNA polymerase cellular reagents were stable for at least 3 weeks when stored at temperatures as high as 42°C (S11 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Isothermal nucleic acid amplification using Bst DNA polymerase cellular reagents.

Indicated copies of synthetic DNA templates derived from human glyceraldehyde 3-phosphate dehydrogenase gene were amplified in LAMP-OSD reactions using 16 units of pure Bst-LF (panel a), 16 units of pure Bst 2.0 (panel b), or 2 x 10⁷ cells of Bst-LF cellular reagents (panel c). Amplicon accumulation was assessed in real time by measuring increase in OSD fluorescence. Representative raw fluorescence amplification curves are depicted in black (6 x 10⁶ template copies), blue (6 x 10⁵ template copies), red (6 x 10⁴ template copies), green (6 x 10³ template copies), pink (600 template copies), purple (60 template copies), yellow (6 template copies), and cyan (0 templates). Cq values obtained using pure commercial Bst-LF (panel a), pure commercial Bst 2.0 (panel b), and Bst-LF cellular reagent (CR) (panel c) are tabulated. Unlike PCR, LAMP is a complex continuous amplification process in which Cq does not always correlate linearly with template copies.

https://doi.org/10.1371/journal.pone.0201681.g006

Cellular reagents for molecular and synthetic biology

Given that cellular reagents were robust in various amplification reactions and formats, we were curious to determine to what extent they could be used in other contexts. Molecular and synthetic biology techniques are currently heavily reliant on the activity of pure enzymes. To demonstrate the possibilities for facile, multi-enzyme cloning procedures with cellular reagents, we created a set of cells that would contain either enzymes or templates as input for one of the most commonly used methods for cloning, Gibson assembly of DNA fragments [24]. Gibson assembly can be used to create vectors from two or more DNA fragments in a one-pot reaction by relying on complementary overlaps created by T5 exonuclease degradation. Once complementary strands have come together, DNA polymerase is used to fill gaps, and Taq DNA ligase seals the remaining nicks to create the new vector.

To most simply carry out Gibson assembly with cellular reagents we merely lyophilized three cell lines that expressed Taq DNA polymerase, Taq DNA ligase, and T5 exonuclease, respectively. We first attempted the simple assembly of a double-stranded gBlock insert (representing a portion of the Aedes albopictus ribosomal protein S7 gene or the E. coli yaiO gene) with 30 bp overlaps corresponding to each end of gel-purified linearized vector, pCR2.1 TOPO (Thermo Scientific) using only cellular reagents. Vector and insert were mixed in equimolar ratio in 1X Gibson assembly buffer and then treated with purified or cellular reagents for Gibson cloning. We hypothesized that endogenous nucleases in BL21 cellular reagents would linearize DNA fragments and prevent recovery of engineered DNA, and therefore opted to use cellular reagents prepared in a Top10 E. coli strain that lacked endonuclease I and recA1. However, initial experiments with these Top10 cellular reagents failed to yield successful DNA assemblies. Optimization experiments revealed that a 10 min pre-heating of Taq DNA polymerase and Taq DNA ligase cellular reagents at 75°C significantly improved Gibson assembly (S2 Table). When DNA parts were mixed with treated cellular reagents comprising 2 x 10⁷ Taq DNA ligase cells, 10⁶ Taq DNA polymerase cells, and 10⁵ T5 exonuclease cells (S2 Table) the Gibson cloning procedure was successful, producing 48–60% of the number of colonies observed with a commercial enzyme master mix containing 0.08 units of T5 exonuclease, 0.5 units of Phusion DNA polymerase, and 80 units of Taq DNA ligase. Background assembly was similar for both cellular and commercial reagents. That said, while most clones assembled by pure enzymes were correct only 50% of the plasmids assembled by cellular reagents had the correct sequence; the remainder were re-circularized vectors.

Despite an increased propensity for incorrect assembly, cellular Gibson assembly is nonetheless a highly enabling technology that stands to significantly reduce costs for and increase accessibility to a core tool for synthetic biology. To further bolster the appeal and utility of cellular Gibson assembly we sought to combine amplification and assembly with cellular reagents. As a first step, cellular Phusion reagents were used to amplify linear vector and insert fragments directly from two different bacterial strains (Fig 7). The kanamycin resistance cassette from the donor plasmid pJLsfGFP and the entire vector backbone of ampicillin-resistant pATetO 6XHis plasmid were amplified with the inclusion of 30 bp overlaps. Following agarose gel purification of the DNAs produced by amplification with cellular reagents, Gibson assembly with either cellular or purified reagents yielded 61 and 884 clones, respectively, that were jointly resistant to both ampicillin and kanamycin; all were verified by sequencing to be correctly assembled (Fig 7). We similarly attempted a three-part assembly with DNAs amplified by cellular reagents by dividing pUC19 into two linear fragments with 30 bp overlaps prior to purification and assembly. The three part assembly yielded 28 clones, all of which were correctly assembled (Fig 7). The improvement in the fraction of correctly assembled vectors in this example (100%) relative to the previous example with gBlock insertion was likely due to the fact that correctly assembled vectors could be directly selected via antibiotic resistance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. PCR and Gibson assembly using cellular reagents.

(a) Schematic depicting cellular PCR followed by cellular Gibson assembly for constructing new plasmids. Bacteria harboring target plasmids are mixed with polymerase-expressing cellular reagents and PCR is initiated by adding appropriate primers, buffer, and dNTP. The resulting PCR products are incubated with cellular reagents expressing Gibson assembly enzymes–Taq DNA polymerase, Taq DNA ligase, and T5 exonuclease–to assemble the new construct. (b) Cellular PCR amplification of vector and insert fragments directly from E. coli bacteria bearing target DNA plasmids using 2 x 10⁷ cells of Phusion cellular reagents. Assembly parts include: (i) “pATetO 6XHis full length” vector for two part assembly with Kan^r cassette bearing appropriate overlapping ends, and (ii) “pUC19 Fragments 1 and 2” for three part assembly with Kan^r cassette whose ends overlap with pUC19 vector fragments. (c) Gibson assembly of agarose gel purified and unpurified cellular PCR products using pure or cellular Gibson assembly reagents. In “negative control” samples the PCR products were incubated in Gibson reaction buffer without pure or cellular Gibson enzymes. “pATetO 6XHis + Kan^r”represents a two part Gibson assembly while “Puc19 Fragment 1 + pUC19 Fragment 2 + Kan^r” represents a three-part Gibson assembly. Representative number of clones recovered in each case in the presence of both ampicillin and kanamycin are reported.

https://doi.org/10.1371/journal.pone.0201681.g007

Finally, we attempted to carry out the same assemblies without DNA purification, using only cellular Gibson reagents. We obtained eight and two colonies, respectively, all of which proved to be the correct assembly (Fig 7).