Week 3 HW: Cell-Free Reaction Design

Part A — Compare cell-free reaction compositions

Each question below shows my own working answer — arrived at with substantial AI research and help — followed by an AI cleaned-up version. See the note on AI use at the bottom.

Part B — publishing 20 µL reaction compositions for sfGFP, PETase, and Reteplase on coopetition.htgaa.org, plus the Community Reagent contribution, is not included in this write-up yet.

Q1 — Why does the concentration of potassium glutamate and magnesium glutamate differ between compositions? Which is more sensitive?

My answer

Short answer

a/ Why they differ. The concentrations differ because (a) they serve and help catalyze different reactions, and (b) different recipes bind differently and require different amounts of K⁺ or Mg²⁺.

b/ Which is more sensitive. Magnesium is more sensitive, as we can see in Table II — it shows that if you overdo the concentration the yield falls off substantially as compared to potassium for the same increase %.

Longer answer

CFPS requires various reagents and buffers to enable the protein generation process to take place, and potassium glutamate and magnesium glutamate both offer various chemical behaviors (providing charged particles) that facilitate that process.

Potassium has one positive charge and magnesium has two — and therefore they have different binding behaviors and serve different purposes.

Imagine trying to fold a long ribbon whose different sections keep pushing one another away.

RNA has a version of that problem: it has many negatively charged parts, which repel one another. Positive particles help soften those electrical pushes. This allows RNA to adopt useful shapes and interact with other molecules. That matters because a ribosome — the machine that assembles protein — is itself made largely of folded RNA and proteins. Potassium helps establish suitable electrical surroundings for these interactions.

Magnesium:

• Keeping the machinery’s shape — it helps stabilize folded RNA and the ribosome.
• Helping chemical reactions happen — many enzymes need magnesium together with ATP.
• Supporting the right molecular connections — the ribosome must attach to molecules, move, and release them as it builds a protein.

When you add magnesium:

• Some associates with RNA, ATP, and other molecules.
• Some remains free in the liquid.
• Changing the ingredients changes that distribution.

For example, PEP binds magnesium more strongly than pyruvate. When the researchers changed the energy ingredient, they therefore retested the magnesium concentration.

For potassium, the authors also tested its concentration within each complete mixture. Its preferred amount depends on the surrounding salts and other ingredients.

So in short, the concentration differs between compositions because of what each of potassium and magnesium binds to — the contents of the solution.

AI cleaned-up answer

1. Why do the concentrations differ between compositions?

Different recipes contain different ingredients and different starting salt levels. Potassium maintains the ionic environment the protein-making machinery needs. Magnesium supports RNA folding, ribosome structure, and every reaction that runs on ATP — but it also binds other components in the mixture, so only part of what you add stays available. Change the ingredients and you change that distribution. For example, PEP binds magnesium more strongly than pyruvate, so switching the energy substrate requires re-tuning magnesium.

2. Which is more sensitive?

Magnesium, clearly, in the formulations tested in Table II:

A comparable percentage increase costs about a third of the yield for magnesium, versus a few percent for potassium. The practical implication: in a screen, magnesium deserves a narrow, finely-spaced range around the optimum, while potassium can be scanned coarsely.

Source: Jewett & Swartz (2004), Table II and accompanying discussion.

Q2 — Explain how a PEP+NTP reaction formulation is different from a Ribose+NMP reaction formulation.

My answer

PEP — helps replenish ATP so energy-dependent tasks can continue.

NTPs — supply the nucleotide form needed to make RNA; ATP and GTP also support energy-consuming tasks.

PEP + NTP: the NTPs supply the building blocks for making RNA from DNA. ATP also powers energy-consuming processes. As ATP is used and ADP is produced, PEP uses that ADP and helps regenerate ATP so the reaction can keep working.

Ribose — a sugar that enzymes in the lysate process to help regenerate ATP from ADP.

NMPs — nucleotides with one phosphate group. Enzymes transfer additional phosphates from ATP onto NMPs, turning them into NTPs with three phosphate groups. RNA polymerase then uses these NTPs to make RNA by following the DNA template.

Ribose + NMP: ribose helps regenerate ATP from ADP. Enzymes use ATP to add phosphates to NMPs, converting them into NTPs. RNA polymerase then uses those NTPs to make RNA from DNA.

Deeper dive

When ATP powers a task, it is chemically changed. Two common outcomes are:

• ATP → ADP — ATP loses one phosphate group.
• ATP → AMP — ATP loses two phosphates together. Attaching an amino acid to tRNA uses this route.

So ADP comes from ATP being used by reactions in the mixture. Some ATP, ADP, and AMP can also arrive with the lysate.

If we keep consuming ATP without replenishing it, energy-dependent steps slow down or stop — even if amino acids and ribosomes are still available. PEP helps turn used ADP back into usable ATP.

AI cleaned-up answer

The two formulations differ in how far upstream they start, and that single choice drives both cost and reaction behavior.

PEP + NTP is the direct route. The nucleotides arrive already in the triphosphate form RNA polymerase needs, and PEP regenerates ATP from ADP in a single enzymatic step. It is fast and predictable, but PEP is consumed stoichiometrically, releases inorganic phosphate that accumulates and inhibits the reaction, and both PEP and NTPs are costly.

Ribose + NMP pushes work onto the lysate’s own metabolism. Ribose is catabolized through the enzymes already present in the extract to regenerate ATP, and that ATP is then used by kinases in the lysate to phosphorylate NMPs up to NTPs before transcription can proceed. More steps have to work, so the formulation depends more heavily on extract quality — but the inputs are dramatically cheaper, and the slower, metabolically-buffered energy release avoids the phosphate accumulation problem.

The trade-off: PEP + NTP buys speed and simplicity with reagent cost; Ribose + NMP buys cost and sustained energy supply by trusting the lysate to do more of the chemistry. This is exactly the trade-off Q3 quantifies.

Q3 — Choose two cell-free reaction compositions from the literature; compare and contrast the choice of reagents and the concentration of reagents.

My answer

Let’s use Olsen (2026) — we can compare PANOx-SP and RF₀ₚₜ.

All concentrations in mM unless otherwise noted.

Like question 2, we have PEP + NTP vs. Ribose + NMP.

We have:

• 2 vs. 5 mM amino acids
• 57 vs. 75 mM HEPES buffer
• 130 vs. 362 mM potassium glutamate

In RF₀ₚₜ we did not add ammonium glutamate. We also did not add spermidine and putrescine. And there are some different metabolic additives between the two.

The RF₀ₚₜ is double the yield at 1% the cost.

We have more protein building material — with the more amino acids. It seems that the engine Ribose + NMP outperformed the PEP + NTP.

AI cleaned-up answer

I compared PANOx-SP (Jewett & Swartz 2004) and RF₀ₚₜ, the cost-optimized formulation, as evaluated by Olsen et al. (2026).

The energy and nucleotide system is the headline difference. As discussed in Q2, PANOx-SP runs on PEP and NTPs; RF₀ₚₜ runs on ribose, glucose, and NMPs.

RF₀ₚₜ also raises several concentrations:

• Amino acids: 2 → 5 mM each — more protein-building material, and a larger buffer against depletion of whichever residue is limiting.
• HEPES: 57 → 75 mM — greater capacity to resist the pH drift that comes with sustained metabolic activity.
• Potassium glutamate: 130 → 362 mM — a substantially different ionic and metabolic environment.
• Magnesium glutamate: unchanged at 8 mM — consistent with Q1: magnesium is the sensitive parameter and sits near its optimum in both.

RF₀ₚₜ also removes components: ammonium glutamate, spermidine, putrescine, tRNA, folinic acid, NAD, CoA, and oxalic acid are all dropped. In their place it adds phosphate (15 mM) and nicotinamide (4 mM). Several of the removed supplements are among the most expensive per-reaction line items, so removal is doing real work on the cost side as well as the chemistry side.

Result:

The complete RF₀ₚₜ formulation therefore produced roughly 2.2× more protein at about 1.5% of the reagent cost — a ~150-fold improvement in cost per gram.

Important caveat: because many ingredients changed simultaneously, this comparison does not establish which individual changes caused the improvement. The formulations were co-optimized as whole systems, and the components likely interact (the elevated potassium, for instance, may only be beneficial in the context of the ribose-driven energy system). Attributing credit would require a component-by-component ablation.

AI use

Consistent with Week 1, here is how AI was used on this assignment.

My answers are substantially supported by AI help. I’m new to biology, and I used AI heavily to get through these papers — to explain the chemistry, to check my understanding, and to help me pull the numbers out of the published formulations. The “My answer” sections are my own working answers in my own words, written after that back-and-forth; they appear here essentially as I drafted them, with spelling corrected and Markdown formatting applied, but the wording and the reasoning are unchanged. The understanding behind them was built with AI, not independently of it.

The “AI cleaned-up answer” sections are exactly what the label says: Claude restructured my answers into tighter prose, pulled the specific numbers out of my working notes into tables, and in a few places added framing I had not written (the screening implication in Q1, the comparison table and phosphate-accumulation point in Q2, the ablation caveat in Q3). I reviewed each one against the sources.

Claude Code handled formatting and publishing — converting my plain-text notes into this Markdown page and committing it to this repository.

References

• Jewett, M. C. & Swartz, J. R. (2004). Mimicking the Escherichia coli cytoplasmic environment activates long-lived and efficient cell-free protein synthesis. Biotechnology and Bioengineering.
• Olsen et al. (2026). Cost-optimized cell-free formulation (RF₀ₚₜ).
• Smith et al. (2026). AI-designed cell-free formulations.