Updated Experimental Plan

Autonomous Bioluminescence in Nicotiana tabacum: Detailed Experimental Plan. Affiliation: Pownall Community Lab, Vancouver BC; HTGAA 2026. Document version: 1.0. Status: working plan, pending Step A verification.

Project framing

The original goal was autonomous firefly bioluminescence in N. tabacum via a five-transcription-unit Golden Gate construct delivering reconstituted luciferin biosynthesis. Literature review has shifted the realistic scope to characterizing the endogenous substrate landscape for autonomous firefly bioluminescence in tobacco, with the minimum viable construct as the primary test and informative-negative outcomes as legitimate deliverables.

The reframe is justified by three findings. First, firefly luciferin chemistry is partly spontaneous (Kanie 2016, Viviani 2022), cysteine and benzoquinone form luciferin without enzyme catalysis. Second, the literature does not document free hydroquinone pools in healthy mature tobacco leaves, making endogenous substrate availability the dominant unknown. Third, glutathione competition kinetics predict benzoquinone half-life on the millisecond scale in plant cells, which combined with low free cysteine (9-10 µM in tobacco per Wirtz 2007) makes spontaneous luciferin formation rate-limited at femtomole-to-low-picomole per gram per hour, four to five orders of magnitude below standard tobacco luciferase imaging thresholds.

The plan therefore emphasizes (a) a sequenced experimental progression with a real deliverable at each milestone, (b) decision points that determine what to do regardless of whether the headline experiment succeeds, and (c) a fallback path that converts a negative minimum-viable-construct result into a substrate-characterization deliverable.

Milestone summary

Total project duration: approximately 4-6 months, fitting within HTGAA 2026 timeline.

Step A: Pre-experimental verification

Objective: Resolve uncertainties that would invalidate downstream work before committing time and reagents to experiments.

Substep A1: Sequence verification

BLAST the ordered PPYR_02911 sequence (Twist synthesis, 1,575 bp, codon-optimized for N. tabacum) against the current Photinus pyralis geneset on i5k Workspace. The official geneset version 1.1 included updated multiple P450 gene models and naming. Confirm that the locus identifier and CDS still match what was ordered. Document the BLAST results with screenshots and save the verified sequence record alongside the Twist order confirmation. Time: 30 minutes.

Decision criterion: if the sequence matches an outdated annotation, re-order before any other Twist orders ship. If the sequence matches the current annotation, proceed.

Substep A2: Construct sequence verification across all three Twist orders

While checking PPYR_02911, also verify TU3 (BGLU46+SKL, 1,590 bp) and TU4 (ACOT9, 1,416 bp) for: (a) absence of internal BsaI sites (Benchling has confirmed for TU2; need to confirm for TU3 and TU4); (b) absence of internal BbsI sites (not previously checked); (c) absence of standard BioBrick sites (EcoRI, XbaI, SpeI, PstI, HindIII, also not previously checked); (d) correct flanking overhangs (CATC at 5′ and GCTT at 3′ for PhytoBricks Level 0 parts). Use Benchling's enzyme finder. Time: 1 hour.

Decision criterion: any flagged internal restriction sites require silent-mutation domestication before Twist ships. If clean, proceed.

Substep A3: HPLC access confirmation

Confirm with Pownall lab whether HPLC is available for substrate measurements, or whether collaboration access exists at UBC, SFU, or BC Cancer. The substrate landscape characterization (Step F primary) and Phase 3 LC-MS confirmation (Step E) both depend on this. Time: depends on response time.

Decision criterion: if HPLC is available or accessible via collaboration, Step F can include direct substrate measurement. If unavailable, Step F shifts entirely to luminescence-readout-only spike-feeding diagnostics, which is workable but produces less mechanistically informative results.

Substep A4: p19 silencing suppressor strain confirmation

Confirm with Michael Rae whether his GV3101 stock includes a p19-expressing strain, or whether p19 needs to be sourced separately. Without p19, transgene expression in tobacco peaks at day 2-3 and crashes by day 5; with p19, expression is sustained through day 7-10. Addgene #74108 is a standard p19 binary vector if a fresh strain is needed.

Decision criterion: if p19 is available, proceed. If not, building or sourcing a p19 strain is added as a Phase 0 sub-task.

Substep A5: Imaging access confirmation

Identify which Vancouver-area lab has IVIS or equivalent cooled CCD imaging access for Phase 2 onwards. Phase 1 with sprayed D-luciferin can use a DSLR with long exposure for qualitative confirmation, but Phase 2 autonomous signal will require proper low-light imaging. Possible options: UBC Centre for Drug Research and Development, UBC Faculty of Medicine core facilities, BC Cancer Research Institute, SFU. Time: 30-60 minutes of email outreach plus response time.

Decision criterion: if imaging access is confirmed, the timeline is unconstrained. If imaging access requires booking weeks in advance or paying per session, build the booking schedule into Phase 2 timing.

Step A deliverable: A one-page verification document listing: PPYR_02911 sequence verification result, all-construct enzyme site clearance, HPLC access status, p19 strain status, imaging facility commitment.

Decision point at end of Step A: If A1 or A2 reveals problems, fix before proceeding. If A3 returns “no HPLC access,” update Step F design to luminescence-only. If A4 returns “no p19,” add p19 strain construction to Phase 0. If A5 returns “no imaging access,” restructure Phase 2 timing around external booking. Otherwise, proceed to Step B.

Step B: Phase 0, RUBY-RED practice infiltration

Objective: Validate the entire infiltration workflow using a published positive control with a visible non-luminescent readout, before committing to luciferase experiments.

Rationale: RUBY (He et al. 2020) produces betalain pigment from tyrosine via three enzymes (CYP76AD1, DODA, glucosyltransferase) that turn infiltrated tissue visibly red. No specialized imaging required, no luciferin substrate, no ambiguity about whether infiltration worked. If RUBY produces visible color, the infiltration technique is sound. If RUBY does not produce visible color, troubleshoot before any luciferase work.

Pre-step preparation (Week 1)

• Grow N. tabacum plants from seed: standard 16-hour day, 22°C, ~5-6 weeks to mature mid-canopy leaves
• Pick fully-expanded leaves 4-6 from the apex on plants showing healthy growth; avoid leaves that are still actively expanding (too fragile, low expression) or starting to senesce (immune system is shifted)
• Acquire RUBY plasmid from Addgene (#160407) and transform into GV3101
• Prepare MMA buffer: 10 mM MES pH 5.6, 10 mM MgCl₂, 100-200 µM acetosyringone (freshly added)

Procedure (Week 2)

• Streak GV3101+RUBY on LB agar with rifampicin + spectinomycin
• Pick single colony, grow overnight in 5 mL LB + appropriate antibiotics, 28°C, 220 rpm
• Pellet at 4000 g for 10 min, resuspend in MMA buffer to OD600 = 0.5
• Incubate at room temperature for 2-3 hours for vir induction
• Co-infiltrate with p19 strain at OD600 = 0.3 (final mixture)
• Use 1 mL needleless syringe pressed against abaxial leaf surface; gently push solution into leaf until visible saturation spreads across infiltration zone
• Mark infiltration zones with permanent marker on the leaf
• Return plant to growth chamber

Readout

• Day 3-4: check infiltration zones daily for visible red coloration
• Day 5-7: photograph infiltration zones; compare to mock-infiltrated zones on same leaf
• Document with date-stamped photos, lighting conditions, and infiltration parameters

Step B deliverable: Photograph of tobacco leaf with visible red RUBY infiltration spots, plus an internal procedural document recording everything that worked and everything that didn't (OD600 used, leaf age, infiltration force, day signal appeared, signal intensity).

Decision point at end of Step B:

• If RUBY produces visible color in infiltrated zones by day 5, infiltration technique is validated. Proceed to Step C.
• If no visible color, troubleshoot before Step C: check Agro viability, check OD600 measurement, check infiltration technique (was solution actually entering the leaf?), check leaf age, check growth chamber conditions. Repeat Step B until validated.
• Common failure modes: leaves too young (still expanding) → use older leaves; OD600 too low → increase to 0.8; infiltration solution running off rather than entering leaf → check syringe technique with experienced colleague.

Step C: TU1 luciferase validation with sprayed D-luciferin

Objective: Validate that luc2+SKL expresses in N. tabacum leaves and targets correctly to peroxisomes, with sprayed exogenous D-luciferin as substrate. This separates “the reporter works” from “the autonomous biosynthesis works” and is a standalone deliverable independent of all subsequent steps.

Rationale: luc2 has been a tobacco reporter since Ow et al. 1986. SKL (Ser-Lys-Leu) C-terminal targeting to peroxisomes was originally characterized in plants. If TU1 fails to produce light when fed luciferin, the construct itself has a problem and downstream experiments are uninterpretable.

Construct preparation

• TU1 assembly via PCR from BBa_K389004 template (luc2 in pSB1C3, iGEM 2021 Plate 4 Well 3C):
• Forward primer: ttttGGTCTCaCCATatggaagatgccaaaaacattaagaagggc
• Reverse primer: ttttGGTCTCaAAGCTTACAGCTTGCTcacggcgatcttgccgcc
• Reverse primer adds SKL (AGC AAG CTG = Ser-Lys-Leu) + TAA stop + GCTT overhang
• High-fidelity PCR (Q5 or Phusion), gel-extract product
• BsaI Golden Gate into Level 1 acceptor pICH47732 with 35S+Ω promoter (BBa_P10100, Plate 4 Well 14D) and NOS terminator (pICH41421, Addgene #50339, replacing BBa_P10401 because the iGEM part uses GCTC overhang instead of GCTT)
• Verify by sequencing with M13F/M13R or vector-specific primers
• Transform verified TU1 Level 1 plasmid into GV3101
• Confirm TU1 strain by colony PCR before glycerol stocking

Co-infiltration components

• Renilla luciferase reporter for normalization: search Addgene for plant-codon-optimized 35S:Rluc8 or 35S:hRluc; if no suitable plasmid exists, build one as a small Golden Gate job using existing 35S+Ω + Rluc8 CDS + a terminator (separate ~2 week sub-project)
• p19 silencing suppressor (separate strain, see Step A4)

Six-spot leaf design (per leaf, 3 plants for biological triplicates)

Total OD600 in each Agro mixture held at 0.8 by adjusting empty-vector co-infiltration, so the only variable across spots 3-6 is TU1 strain proportion.

Imaging protocol (day 5 post-infiltration)

• Detach or image attached leaves in the dark
• Spray 1 mM D-luciferin in 0.01% Triton X-100 across abaxial surface
• Wait 5 minutes for substrate uptake
• Image firefly luminescence: open filter or 560 nm bandpass, 30 second to 5 minute exposure depending on signal strength
• Wait 30+ minutes for D-luciferin signal decay or wash with PBS
• Spray 50 µM native coelenterazine in PBS
• Image Renilla luminescence: 480 nm bandpass, similar exposure
• Quantify total photon flux per spot using ROI software (Living Image, Fiji)
• Compute firefly/Renilla ratio per spot

Western blot confirmation (in parallel)

• Extract total protein from one leaf disc per spot at day 5
• Western with anti-firefly luciferase antibody (Promega G7451)
• Loading control: anti-actin or anti-tubulin
• Confirms expression independent of activity

Step C deliverable: Imaged tobacco leaf with quantified luminescence at TU1 spots above mock and empty-vector backgrounds, optimal OD600 documented, Western blot confirming luc2 expression.

Decision point at end of Step C:

• If TU1 produces light significantly above background AND Western shows expression: reporter validated. Proceed to Step D using the optimal OD600 from this round.
• If TU1 produces light but Western is weak/absent: technical artifact in Western, repeat. Proceed to Step D anyway since the functional assay is the priority.
• If no luminescence but Western shows expression: protein is made but not active. Possible issues: SKL targeting failure (try cytosolic luc2 in parallel), substrate access (try vacuum infiltration of luciferin instead of spray), construct error (re-sequence). Do not proceed to Step D until resolved.
• If no luminescence and no Western signal: construct or expression problem. Re-sequence TU1, check 35S+Ω promoter integrity, check NOS terminator, repeat infiltration with fresh Agro stock. Do not proceed.

Step D: TU1 + TU5 minimum viable construct test

Objective: Test whether co-infiltration of luc2+SKL plus AtLAC17 produces detectable autonomous bioluminescence in tobacco without sprayed luciferin substrate. This is the core experimental hypothesis of the reframed project.

Hypothesis: AtLAC17 oxidizes endogenous tobacco hydroquinones to benzoquinones; benzoquinones react spontaneously with endogenous tobacco cysteine to form L-luciferin, which racemizes to D-luciferin via endogenous esterases, CoA, and luciferase activity (per Viviani 2022 mechanism); luc2+SKL produces light from D-luciferin.

Realistic expected outcome (per substrate landscape literature review): Most likely outcome is non-detectable or barely-above-background signal due to (a) lack of documented free hydroquinone pools in healthy tobacco; (b) compartment mismatch between apoplastic AtLAC17 and intracellular cysteine; (c) glutathione quenching of benzoquinone on millisecond timescale at millimolar GSH concentrations; (d) low free cysteine (9-10 µM) competing against millimolar GSH for the same electrophile. A negative result is informative and feeds directly into Step F.

Construct preparation

• TU5 (AtLAC17): PCR or synthesize Arabidopsis LAC17 CDS (At5g60020); standard PhytoBricks flanking; PCR from cDNA or order from Twist; consider whether to use AtLAC17 with native signal peptide (default apoplastic targeting) vs. variants without signal peptide (cytosolic, but laccases require specific maturation that may not occur outside the secretory pathway)
• BsaI Golden Gate into Level 1 acceptor with 35S+Ω promoter and a terminator distinct from TU1's NOS (use OCS or 35S terminator)
• Sequence verify, transform into GV3101

Six-spot leaf design (per leaf, 3 plants)

Imaging protocol (day 5 post-infiltration)

• Image firefly luminescence at the longest exposure the camera supports (typically up to 5 minutes for cooled CCD systems)
• Image Renilla after coelenterazine spray for normalization
• For TU1+TU5 spots specifically, do not spray any substrate before the firefly imaging, these spots are testing whether endogenous chemistry is sufficient
• Quantify firefly/Renilla ratio per spot
• Compare TU1+TU5 spots to mock and empty-vector spots for above-background signal

Step D deliverable: Imaged leaf with quantified TU1+TU5 luminescence vs. controls, with explicit statement of whether signal is above background and at what magnitude.

Statistical threshold for “above background”: TU1+TU5 firefly/Renilla ratio significantly greater than empty-vector firefly/Renilla ratio (paired t-test or mixed-effects model, p < 0.05 with biological triplicates), AND fold-change of at least 2x over background. This threshold is set conservatively because plant tissue auto-luminescence is real and stress-dependent.

Decision point at end of Step D:

• If TU1+TU5 produces signal above background per the threshold above: minimum viable hypothesis validated. Proceed to Step E (optimization).
• If TU1+TU5 does not produce signal above background: substrate supply is limiting. Proceed to Step F (substrate characterization).
• If TU1+TU5 produces inconsistent signal (some leaves yes, some no): proceed to Step F's spike-feeding diagnostics to characterize what's variable.

Step E: Optimization path (if Step D positive)

Objective: Maximize and characterize autonomous bioluminescence signal; integrate additional TUs as optional enhancements; produce manuscript-quality documentation of the system.

This path is the success-case scenario. Each substep is an independent ~10-day experimental cycle, not a dependency chain. Order based on which questions are most informative, not which are easiest.

Substep E1: OD600 ratio optimization between TU1 and TU5 strains

Vary the proportion of TU1 to TU5 in the Agro mixture (1:1, 1:2, 2:1, 1:5, 5:1) at fixed total OD600. Identifies whether the system is rate-limited by luciferase enzyme abundance or substrate generation. ~10 days.

Substep E2: Time course

Image TU1+TU5 spots at days 3, 5, 7, 10 post-infiltration. Characterizes signal kinetics, peak time, decay rate. Important for distinguishing “weak but sustained” from “strong but brief” expression patterns.

Substep E3: Tissue extraction and LC-MS confirmation (if HPLC accessible)

Extract day-5 infiltrated tissue, run LC-MS targeting D-luciferin (m/z 281.0049 [M+H]+, 278.9904 [M-H]-) with chiral separation if possible. Confirms that the signal is genuine luciferase-on-luciferin chemistry and not autoluminescence. If LC-MS is not accessible, this substep is replaced with biochemical confirmation: extract leaf, add to recombinant luciferase + ATP in vitro, observe whether extract supports luminescence (should show D-luciferin presence).

Substep E4: Add TU3 (BGLU46+SKL)

Test whether glucoside mobilization improves signal. TU1+TU5+TU3 vs. TU1+TU5. If signal increases significantly with BGLU46, tobacco contains usable hydroquinone glucoside pool; if no change, the assumption was wrong.

Substep E5: Add TU2 (PPYR_02911)

Test whether oxidative tailoring improves signal. TU1+TU5+TU2 vs. TU1+TU5. Note: per the literature review, this may require also co-expressing P. pyralis or insect cytochrome P450 reductase for activity. If adding PPYR_02911 alone produces no change, repeat with co-expressed insect CPR before concluding the candidate is inactive. Cloning P. pyralis CPR from Fallon transcriptome data is a separate ~3 week sub-project if PPYR_02911 alone doesn't help.

Substep E6: Add TU4 (ACOT9)

Test whether stereochemistry conversion improves signal. TU1+TU5+TU4 vs. TU1+TU5. If signal increases, the L-to-D racemization step is rate-limiting in the basic system; if no change, racemization is happening adequately via endogenous mechanisms.

Substep E7: Combinatorial test of the most-effective additions

Based on E4-E6 results, build the combination that produced the strongest signal and characterize it fully (time course, OD600 optimization, tissue distribution, signal duration over multiple imaging sessions).

Step E deliverable: Manuscript-quality characterization of autonomous bioluminescence in tobacco with the identified optimal construct. Components: photograph of glowing tobacco leaf as headline figure, OD600 optimization curve, time course, LC-MS or biochemical luciferin confirmation, contribution analysis of optional TUs (TU2, TU3, TU4 each tested for impact). This is publishable.

Decision point at end of Step E: Project meets HTGAA deliverable. Continue toward stable transformation (Phase 4, post-HTGAA) using the optimized construct.

Step F: Substrate characterization path (if Step D negative)

Objective: Identify which substrate, compartment, or competing reaction is limiting autonomous bioluminescence in tobacco. Convert a negative minimum-viable-construct result into a substrate-landscape characterization deliverable.

Path F-A: With HPLC access (preferred)

Substep FA1: Direct measurement of free substrate pools

Extract leaf tissue from mock-infiltrated, empty-vector-infiltrated, and TU1+TU5-infiltrated leaves at day 5. LC-MS or HPLC-UV targeting:

• Free hydroquinone (C₆H₆O₂, 110.0368 monoisotopic, 109.0295 [M-H]⁻)
• Free catechol (same mass as hydroquinone, separable by retention time)
• Arbutin (C₁₂H₁₆O₇, 272.0896, 271.0823 [M-H]⁻)
• Free cysteine (C₃H₇NO₂S, 121.0198, 122.0270 [M+H]⁺), likely needs HILIC or derivatization for retention
• Free glutathione (C₁₀H₁₇N₃O₆S, 307.0838, 306.0765 [M-H]⁻)

Outcome: documented quantitative substrate landscape in the actual experimental conditions. This is the primary substrate-characterization deliverable.

Substep FA2: Spike-feeding rescue panel

Co-infiltrate TU1+TU5 spots with various exogenous substrates and image:

• Spot A: TU1 + sprayed D-luciferin (positive control)
• Spot B: TU1 + TU5, no spike (negative control)
• Spot C: TU1 + TU5 + 100 µM hydroquinone co-infiltrated
• Spot D: TU1 + TU5 + 1 mM L-cysteine co-infiltrated
• Spot E: TU1 + TU5 + 100 µM hydroquinone + 1 mM L-cysteine co-infiltrated
• Spot F: TU1 + TU5 + 10 µM benzoquinone (low dose, toxicity check first)

Pattern interpretation:

• Spot C rescues, others don't → hydroquinone supply is the limit
• Spot D rescues, C doesn't → cysteine access is the limit
• Spot E rescues, neither C nor D alone → both are limiting in combination
• Spot F rescues, C doesn't → AtLAC17 oxidation step is the limit (laccase isn't producing benzoquinone fast enough or in the right place)
• Nothing rescues except Spot A → either GSH quenching is dominant or the spontaneous chemistry isn't happening at all in plant cells

Substep FA3: Compartment intervention test

Build a cytosolic luc2 variant (TU1 minus SKL, single primer change on the reverse primer). Test TU1-cytosolic + TU5 vs TU1-peroxisomal + TU5. If cytosolic luc2 produces signal where peroxisomal doesn't, the compartment mismatch is real and the engineering solution is laccase relocalization or luciferase relocalization, not biosynthesis enzyme addition.

Substep FA4: GSH modulation test

Use buthionine sulfoximine (BSO), a γ-glutamylcysteine synthetase inhibitor that depletes cellular GSH, applied to leaf tissue prior to imaging. If BSO treatment rescues TU1+TU5 signal, GSH quenching is the dominant failure mode. If BSO has no effect, GSH is not the bottleneck. Caveat: BSO is broadly toxic to plant tissue; this is a perturbation experiment, not a sustainable engineering solution.

Path F-B: Without HPLC access (fallback)

If Step A determined HPLC is not available, Step F runs with luminescence-only readouts. Skip FA1 entirely. FA2 (spike-feeding) becomes the primary diagnostic. FA3 and FA4 proceed as designed since they use luminescence readouts.

In this case, the Step F deliverable shifts from “quantitative substrate landscape characterization” to “rescue-pattern-based identification of the limiting factor.” Less mechanistically informative but still a real deliverable.

Step F deliverable: Documented identification of which substrate, compartment, or competing reaction is limiting autonomous firefly bioluminescence in tobacco. Components: substrate pool measurements (with HPLC) or rescue pattern (without), interpretation of which engineering intervention is most likely to move the system from non-detectable to detectable, recommendations for staged Phase 4+ work. This is publishable as a foundational characterization paper for the field, framed as “what would have to be true for autonomous firefly bioluminescence in plants to work.”

Decision point at end of Step F: Project meets HTGAA deliverable. Recommendations from F feed into post-HTGAA engineering work, different recommendations depending on what the limiting factor turned out to be:

• Hydroquinone-limited → upstream phenylpropanoid pathway engineering (PAL boost, dedicated hydroquinone synthase) or genetic substrate supplementation
• Cysteine-limited → SAT/OASTL overexpression or other sulfur metabolism intervention
• Compartment-limited → laccase relocalization (cytosolic AtLAC17 variant, or different laccase with native cytosolic expression)
• GSH-quenching-limited → harder problem; possibly requires apoplast-targeted reaction with apoplast-stable cysteine analog, or transient GSH knockdown via RNAi
• Multiple factors limiting → staged engineering with each factor addressed sequentially

Step G: Project handoff

Objective: Document everything in a form that supports continuation of the work post-HTGAA, whether by me or by future collaborators / employees.

Components:

• Full experimental record in lab notebook format
• Sequence files for all constructs in Benchling and SnapGene formats
• Glycerol stocks of all GV3101 strains, properly labeled and stored at -80°C
• Plasmid stocks of all Level 0, Level 1, and Level 2 constructs
• Photographic record of every infiltration experiment with metadata
• Quantitative imaging data exports (raw photon flux per spot) with ROI definitions
• Western blot images and quantifications
• Metabolomics data (if HPLC accessible)
• Final report: 10-15 page document covering the full project arc, results, and forward roadmap
• Foxfire/Lucera/Kepler-relevant artifact: distilled summary appropriate for investor or hiring conversations

Step G deliverable: Project handoff package, suitable for HTGAA final presentation and for continuation into post-HTGAA work.

Risk register

Things that could go wrong, in roughly decreasing order of likelihood:

Risk 1: Step C fails (TU1 doesn't produce light). Likelihood: low to moderate. Mitigation: Western blot identifies whether it's expression or activity; troubleshooting decision tree exists. Project pauses but doesn't restart from zero.

Risk 2: Step D negative result is “ambiguous” rather than clearly negative. Likelihood: moderate. Some signal slightly above background but not statistically significant, or inconsistent across leaves. Mitigation: Step F spike-feeding is designed to handle this; ambiguous Step D becomes “characterize the variability” project.

Risk 3: Twist orders contain undetected internal restriction sites. Likelihood: low after Step A2 verification. Mitigation: Step A catches this. If discovered after shipping, silent-mutation domestication via Q5 site-directed mutagenesis is feasible (~2 weeks per construct).

Risk 4: Imaging access becomes a bottleneck. Likelihood: low to moderate, depends on Step A5 outcome. Mitigation: book imaging slots well in advance; have DSLR backup for qualitative confirmation; build infiltration timing around imaging availability rather than the reverse.

Risk 5: HPLC access turns out to be unavailable. Likelihood: moderate, depends on Step A3. Mitigation: Step F has a Path F-B that runs without HPLC; the deliverable degrades but doesn't disappear.

Risk 6: Plant supply runs out. Likelihood: low if seed-starting is ongoing. Mitigation: stagger plant seeding so new mature plants are available every 2 weeks.

Risk 7: Agrobacterium contamination kills cultures. Likelihood: low with good technique. Mitigation: maintain glycerol stocks; re-streak from -80°C any time a culture seems off.

Risk 8: GV3101 strain loses p19 plasmid over passages. Likelihood: low to moderate over many passages. Mitigation: re-streak from glycerol stock for each major experiment; verify p19 by colony PCR if signal seems to drop.

Risk 9: I run out of time or energy. Likelihood: real and worth naming. Mitigation: the phased plan produces a deliverable at every milestone, so partial completion is still meaningful project work. If life intervenes after Step C, Step C alone is a real result.

Open questions and assumptions

These are things this plan assumes that haven't been verified, listed so they can be revisited if any turn out to be false:

• That Pownall lab has bench space and equipment time available throughout the project duration
• That GV3101 + p19 with Michael Rae's protocols transforms reliably into N. tabacum
• That tobacco plants are available and growing at sufficient rate to keep up with experimental cadence
• That AtLAC17 expressed under 35S in tobacco produces functional, properly-folded protein with copper loading (not previously demonstrated in tobacco specifically; Arabidopsis native expression context may not transfer cleanly)
• That Renilla luciferase normalization works in tobacco at the substrate concentrations chosen (50 µM coelenterazine; some plant tissue is reported to interfere with coelenterazine but tobacco should be acceptable per Curtis 2025)
• That HPLC access, if confirmed in Step A3, includes appropriate columns and method development support, not just instrument access

What this plan is not

This plan is not an attempt to actually deliver autonomous bioluminescent tobacco within HTGAA 2026. The literature predicts that's unlikely to happen on the timeline available with the substrate constraints documented. This plan is instead a sequenced experimental progression that produces a real deliverable at each milestone, characterizes the substrate landscape that determines what would need to change for the original goal to be achievable, and stages the engineering work for post-HTGAA continuation. The original 5-TU construct is preserved as the post-HTGAA optimization target rather than the HTGAA deliverable.

The honest scope statement: HTGAA delivers either (a) the first autonomous firefly bioluminescence in plants with a 2-enzyme construct, or (b) the first quantitative characterization of why endogenous tobacco substrates are insufficient for autonomous firefly bioluminescence and what would be needed. Both are worth doing. The experiment determines which path the project takes.

Document end. Version 1.0, drafted as a working plan pending Step A verification. Will be updated as Step A outputs come in.