Cycle 1: Design (D)

Parts / Circuit Design

The target gene is an anti-IL-23 scFv (ustekinumab-derived) arranged as VH–(G4S)₃–VL. To enable secretion, we fused an N-terminal SlpA signal peptide; a C-terminal 6×His tag facilitates detection and downstream purification. The expression backbone is pLEM415 under the ldhL promoter. Practically, we started from the pLEM415-ldhL-mRFP1 plasmid and commissioned removal of the mRFP1 ORF, replacing it with a codon-optimized anti-IL-23 scFv while leaving regulatory elements unchanged—thereby minimizing confounders and preserving comparability with existing workflows. Use of pLEM415/ldhL in Lactobacillus expression is documented and materials are publicly available.

Rationale and Tools

Immunoglobulin variable domains require correct disulfide bond formation to maintain structure and antigen binding; directing the product through the Gram-positive Sec pathway and into the extracellular milieu can favor proper folding and reduce intracellular aggregation. Ustekinumab binds the shared p40 subunit of IL-12/IL-23, blocking receptor engagement; leveraging its variable regions provides biochemical specificity for the IL-23 axis. At the sequence level, we performed codon optimization for Lactobacillus and employed standardized cloning overhangs for modular assembly; plasmid maps and junctions were planned with routine molecular design software.

Sequence Disclosure (for Reproducibility)

Provide the complete sequences below to enable exact replication. Include both nucleotide (codon-optimized for Lactobacillus) and amino-acid sequences, with clear domain boundaries:

Construct:pLEM415-ldhL-SlpA–anti-IL-23 p40 scFv–6×His

Topology:SlpA SP – VH – (G4S)₃ – VL – 6×His

IP / Ethics Notice

The ustekinumab-derived variable regions used in this project are for academic research and educational purposes only. Related sequences/uses may be protected by patents and trademarks held by Janssen Biotech/Johnson & Johnson. Any applications beyond research (e.g., commercialization, clinical use) are outside the scope of this project and require prior authorization from the rights holders.

Cycle 1: Build (B)

Manufacturing

The construct will be built by GentleGen according to our specifications: starting from pLEM415–Idhl–mRFP1, using ApaI (5′) and ClaI (3′) restriction sites to remove the mRFP1 ORF and replace it with a codon-optimized SlpA–VH–(G4S)₃–VL–6×His insert (optimized for Lactobacillus). All other regulatory elements (Idhl promoter/RBS/terminator) remain unchanged.

Complete annotated map

Cycle 1・TEST

Sequence QC (TEST)

Vendor and in-house verification identified two variants in the scFv CDS: a 1-bp deletion of G at nt 755 and a C→T transition at nt 1163.By aligning Sanger reads (CX9047_C10.pdf, CX9048_E07.pdf, M13F-77_C09.pdf) to the codon-optimized reference (A243050-1full.pdf) using NCBI BLASTn, we confirmed both events. Results revealed a 1-bp G deletion at nt755, producing a frameshift and predicted premature stop, which compromises downstream domains and likely disrupts the C-terminal 6×His tag. We also observed a C→T substitution at nt1163; its protein-level consequence depends on codon context (synonymous vs. missense vs. nonsense) but does not alter the decision once the frameshift is present. Due to the observed deletion mutation (1-bp deletion causing a frameshift), we will not proceed with functional assays for this construct.

Sequence QC (Evidence)

To support these findings, we provide BLAST alignment snapshots and Sanger chromatograms:

1. BLAST

   

   Alignment snapshot around the predefined coordinate (nt755) with Query = Sanger read and Subject = codon-optimized reference CDS. The dash (–) indicates a single-base gap relative to the reference, consistent with a 1-bp deletion in the insert that would shift the reading frame and predict a downstream premature stop.

   

   Alignment snapshot around the predefined coordinate (nt1163) with Query = Sanger read and Subject = codon-optimized reference CDS. A single base differs between the two sequences (C→T in the read), consistent with the reported SNP at nt1163.

2. Sanger

   

   

Cycle 1・LEARN

Observations

During E. coli amplification, the insert repeatedly accumulated InDels (including a 1-bp deletion at nt755) and point mutations, preventing recovery of an intact construct. The issue persisted even with low-recombination strains, indicating sequence intrinsic and context-dependent instability.

Hypothesized Causes

Sequence-level: the scFv cassette likely contains homopolymers and/or micro-repeats, promoting polymerase slippage and recombination.

Context effects: within the pLEM415–IdhL framework, local transcriptional read-through, long recombination-prone stretches, or expression burden may increase instability.

Decision

To avoid further delays, we will pivot to a mammalian system for functional validation while re-engineering the bacterial construct for stability, then back-port it to Lactobacillus once hardened.

Expansion into Cycle 2 (broader candidate pool)

To reduce single-sequence risk and enrich the study, we added three de novo scFv candidates (1/2/3) plus a risankizumab-derived anti-IL-23 p19 scFv and an ustekinumab-derived anti-IL-23 p40 scFv, all expressed from pTwist-CMV (mammalian) for the second round. This covers p19 (risankizumab-derived) and p40 (ustekinumab-derived) epitopes, improving our odds and widening optimization space.

Reference

1. del Rio, B., Redruello, B., Fernandez, M., Martin, M. C., Ladero, V., & Alvarez, M. A. (2019). Lactic Acid Bacteria as a Live Delivery System for the in situ Production of Nanobodies in the Human Gastrointestinal Tract. Frontiers in Microbiology, 9, 3179.
2. Bao, S., Zhu, L., Zhuang, Q., Wang, L., Xu, P. X., Itoh, K., Holzman, I. R., & Lin, J. (2013). Distribution dynamics of recombinant Lactobacillus in the gastrointestinal tract of neonatal rats. PloS one, 8(3), e60007.
3. Klotz, C., & Barrangou, R. (2018). Engineering Components of the Lactobacillus S-Layer for Biotherapeutic Applications. Frontiers in microbiology, 9, 2264.

Cycle 2・Design (D)

Parts / Circuit Design

The targets for Cycle 2 are five secreted scFv candidates expressed in HEK293T: three de novo scFv (1/2/3), a risankizumab-derived anti-IL-23 p19 scFv and an anti-IL-23 p40 scFv. Each construct follows the same topology Igκ leader–VH–(G4S)₃–VL–6×His, driven by the CMV promoter on a pTwist-CMV backbone with a standard polyA signal. Human codon optimization is applied; the C-terminal 6×His tag supports detection (anti-His Western) and optional Ni-NTA cleanup.

Rationale and Tools

Secretion via the Igκ signal peptide routes the scFv through the ER/Golgi for oxidative folding and disulfide formation, improving the chance of functional paratopes relative to bacterial cytosol. HEK293T enables rapid, high-level transient expression to generate clean supernatants for binding/neutralization assays. Design steps include human-preferred codons and standardized junctions to keep modules swappable. Plasmid maps and junctions are planned with routine tools, keeping all candidates in a single, comparable expression context (CMV/Igκ/(G4S)₃/His).

Sequence Disclosure (for Reproducibility)

Provide full nucleotide (human codon optimized) and amino-acid sequences with labeled domain boundaries.

• Construct(s):
  1. pTwist-CMV–Igκ–scFv 1–6×His
  2. pTwist-CMV–Igκ–scFv 2–6×His
  3. pTwist-CMV–Igκ–scFv 3–6×His
  4. pTwist-CMV–Igκ–anti-IL-23 p19 scFv(risankizumab-derived)–6×His
  5. pTwist-CMV–Igκ–anti-IL-23 p40 scFv(ustekinumab-derived)–6×His
• Topology: Igκ SP – VH – (G4S)₃ – VL – 6×His – Stop codon

IP / Ethics Notice

Any ustekinumab-derived and risankizumab-derived variable regions are used for academic research and education only. Related sequences/uses may be protected by the rights holders. Applications beyond research (e.g., commercialization, clinical use) are out of scope and require prior authorization.

Cycle 2: Build (B)

Manufacturing

The scFv CDS (human codon optimized) will be synthesized and assembled by Twist Bioscience directly into the pTwist-CMV backbone downstream of the Igκ signal peptide, following the shared topology Igκ–VH–(G4S)₃–VL–6×His. All regulatory elements (CMV promoter/leader, polyA/terminator, selection marker) remain unchanged to minimize confounders and enable fair comparison across candidates.

Complete annotated map

1. pTwist-CMV–Igκ–scFv 1–6×His
   
2. pTwist-CMV–Igκ–scFv 2–6×His
   
3. pTwist-CMV–Igκ–scFv 3–6×His
   
4. pTwist-CMV–Igκ–anti-IL-23 p19 scFv(risankizumab-derived)–6×His
   
5. pTwist-CMV–Igκ–anti-IL-23 p40 scFv(ustekinumab-derived)–6×His
   

Cycle 2: Test (T)

Objectives

1. Determine whether each scFv is secreted by HEK293T (CMV–Igκ–VH–(G4S)₃–VL–6×His).
2. Assess binding to IL-23 by ELISA.

Methods (concise)

• Expression: pTwist-CMV constructs; transient HEK293T (48–72 h).
• Samples: collect supernatant and cell lysate separately.
• Western blot: reducing SDS-PAGE; anti-His detection; equal loading and identical exposure across lanes.
• ELISA: IL-23 coated plates; serial dilutions of supernatant/lysate; anti-His-HRP detection; background-subtracted.

Results

1. Secretion & Binding Overview

   anti-IL-23 p19 scFv and anti-IL-23 p40 scFv show robust secretion and binding, while anti-IL-23 scFv-1/2/3 are expressed intracellularly (bands at ~25 kDa in lysate) but not secreted. Supernatant Western blots display clear bands at ~50 kDa for p19/p40 and none for #1/2/3; lysate Westerns show ~50 kDa bands for p19/p40 and ~25 kDa bands for #1/2/3, indicating expression without secretion for the scFv-1/2/3. Consistently, supernatant ELISA curves are dose-dependent for p19/p40 (p40 strongest), whereas scFv 1 remains in the background. Lysate ELISA confirms p19/p40 retain IL-23 binding activity in cell extracts; scFv 1/2/3 stay near baseline.

2. Figure

   

   Strong bands at ~50 kDa for p19 and p40 indicate efficient secretion; scFv-1/2/3 show no detectable supernatant bands. Reducing conditions; equal loading; identical exposure.

   

   p19/p40 show bands at ~50 kDa; scFv-1/2/3 show intracellular bands at ~25 kDa, confirming expression but secretion bottlenecks.

   

   Dose–response curves: p40 strongest (OD450 ≈ 1.2 at 20 µg), p19 positive (≈ 0.7), scFv-1 at background. Same plate/conditions; background-subtracted.

   

   Dose–response curves: p40 (≈ 1.15 at 50 µg) and p19 (≈ 0.8 at 50 µg) remain positive; scFv-1/2/3 near background; NCV baseline. Background-subtracted; same detection.

What we will do next

p19-scFv & p40-scFv — proceed to functional characterization

Pathway inhibition: Run IL-23–induced pSTAT3 assay (reporter or phospho-flow) on Caco-2 cell model.

scFv-1 / scFv-2 / scFv-3 — collaborate with biomedical company

External production confirmation by Leadgene Biomedical:After reaching out, Leadgene Biomedical has successfully expressed scFv-1/2/3 in HEK293. We will document their expression proof (report/traces) on the LEARN part. This external success indicates that the sequences are producible in HEK cells and supports the hypothesis that our in-house non-secretion stems from chassis rather than an absolute misfolding defect.

Cycle 2: Learn (L)

What we observed (quantitative summary)

• Secretion vs. expression. In HEK293T (adherent), p19/p40 produced clear supernatant bands (~50 kDa) and dose-dependent IL-23 binding. scFv-1/2/3 showed lysate bands only (~25 kDa) and no supernatant signal, with ELISA near the background.
• External replication. Leadgene reported successful HEK293 (suspension) expression of scFv-1/2/3 under their platform, indicating that these sequences are expressible/secreted in a different 293 context.

Why HEK293T (adherent) failed while HEK293 (suspension) worked — model and literature

1. Host format & medium: Suspension-adapted HEK293 variants (e.g., HEK293F/293-3F6) in serum-free, chemically defined media routinely achieve higher titers of secreted proteins and antibodies in transient runs, owing to higher viable cell densities and media/additive optimization for secretion and oxidative folding. Multiple studies document robust, scalable transient production of secreted proteins/antibodies in suspension HEK293 systems.
2. Lineage differences: HEK293T carries SV40 large T antigen and is widely used for transient expression, but suspension 293 derivatives are frequently optimized for secretory outputs; depending on construct context (leader/order/linker/tag), secretion can be superior in suspension lines despite strong intracellular expression in 293T. Comparative and review articles note distinct behaviors among HEK293 sublines (293, 293T, 293F, 293E), including secretion performance and bioprocess suitability.
3. Adherent↔suspension rewiring: Global transcriptomic/proteomic analyses show systematic differences between adherent and suspension HEK293, including lipid/cholesterol metabolism and secretory pathway components—factors that can modulate ER-to-Golgi trafficking, glycosylation, and secretion efficiency. This provides a mechanistic basis for “same CDS, different secretion outcome” across hosts.
4. Antibody formats in 293 systems: Transient antibody/scFv formats have been repeatedly optimized in HEK293 backgrounds (including scFv-Fc and secreted domains), supporting the general claim that 293 suspension platforms are suitable antibody hosts.

Reference

1. Jäger, V., Büssow, K., Wagner, A., Weber, S., Hust, M., Frenzel, A., & Schirrmann, T. (2013). High level transient production of recombinant antibodies and antibody fusion proteins in HEK293 cells. BMC biotechnology, 13, 52.
2. Malm, M., Saghaleyni, R., Lundqvist, M. et al. Evolution from adherent to suspension: systems biology of HEK293 cell line development. Sci Rep 10, 18996 (2020).
3. Lao, T., Farnos, O., Bueno, A., Alvarez, A., Rodríguez, E., Palacios, J., Luz, K. R. d. l., Kamen, A., Carpio, Y., & Estrada, M. P. (2023). Transient Expression in HEK-293 Cells in Suspension Culture as a Rapid and Powerful Tool: SARS-CoV-2 N and Chimeric SARS-CoV-2N-CD154 Proteins as a Case Study. Biomedicines, 11(11), 3050.