Introduction
Our project aimed to explore the feasibility of producing intrinsic factors from human, bovine, rat, platypus and porcine and to assess their potential use in oral vitamin B12 supplementation. By combining wet lab experiments with dry lab modelling, we demonstrate a full pipeline from gene design to functional assays, compare cross-species differences, and identify the most promising candidates for future development.
Achievement 1: Successful Construction of Expression Vectors
What we achieved:
We designed, cloned and constructed expression constructs for expressing IFs from five species using the pGAPZα A expression system. The genes were synthesized with adapter sequences (see Experiment 1) and cloned following PCR amplification (see Experiment 2), restriction digestion (see Experiment 4), and ligation into dephosphorylated vectors (see Experiment 5).
Our critical discussion:
This achievement validated successful construction of the expression vectors used for expressing IFs in Komagataella phaffii. The colony PCR (figure 1, see Experiment 6) identified positive E. coli transformants containing the insert and with sequencing (see Experiment 7) we were able to confirm the structural integrity of at least one construct per IF variant. Although colony PCR identified multiple positive colonies, finding positive colonies for the human IF construct proved to be particularly challenging. Additionally, the sequencing results indicated that a few mutations had occurred during expression vector construction. The chosen assembly method, restriction cloning, proved to be successful, but it can be risky and time-consuming, highlighting the need to consider alternative assembly methods in the future, such as Golden Gate or Gibson Assembly.
Figure 1. Representative colony PCR confirming correct IF inserts across species. Rat and porcine colonies yielded higher DNA concentrations; human IF required additional colony screening.
Achievement 2: Expression of IFs using Komagataella phaffii
What we achieved:
We successfully expressed all IF variants using Komagataella phaffii and the pGAPZα A expression system. Multiple colonies were screened for IF expression for each variant and through expression monitoring using SDS-PAGE we were able to identify the colonies with the highest IF expression (see Experiment 10).
Our critical discussion:
Our initial attempts at expressing IFs proved to be unsuccessful due to contamination in the expression cultures. However, after introducing Zeocin to the expression media and ensuring sterile working practices, we were able to successfully express IF variants from all species. Screening multiple colonies for IF expression and monitoring the expression over several days revealed marked differences between colonies. For bovine and rat IF, we identified colonies producing strong, distinct ~50 kDa bands, consistent with high yield and good secretion efficiency. In contrast, colonies expressing human and platypus IF produced only faint bands, suggesting lower expression or secretion losses. For porcine IF, moderate expression was observed, but as we will explain in following achievement steps in this proof-of-concept section, it was later excluded after SEC revealed no recoverable protein.
Although we were able to express IFs from all species and identify colonies with the best expression, the expression yields remained fairly moderate for some variants. This highlighted the need for future expression optimization and additional screening of colonies for finding the optimal strategies for production on a larger scale. When discussing the observed differences between species, we hypothesized that species‑specific folding requirements, glycosylation patterns, or sensitivity to yeast proteases, might have an impact in addition to the differences originating from the chosen colonies. Interestingly, dry lab MD simulations (see Dry Lab) predicted bovine IF to have a more stable, compact fold under simulated aqueous conditions, which may partly explain its high yield and clean SDS‑PAGE profile.
Figure 2. A) SDS‑PAGE screening identified high‑yield expression cultures for each species. In particular, bovine and rat IF showed strong ~50 kDa bands, B) Overlay chart of dry lab RMSD plots for each species, emphasizing bovine’s compactness.
Achievement 3: Purification and His-tag Removal for Native IF Production
What we achieved:
We purified IF proteins via Ni‑NTA affinity chromatography and removed His‑tags using TEV protease (see Experiment 12, Experiment 11).
Our critical discussion:
Purification yielded species‑dependent differences in both quantity and purity. Rat and bovine IF again emerged as high‑yield candidates, with clear post‑TEV bands near 45-50 kDa. Human and platypus IF were present but at lower concentrations, and porcine IF recovery was minimal. At this point, we hypothesised that lower yields for human and platypus IF may partly result from nonspecific losses during concentration and potential aggregation during Ni‑NTA elution, especially if glycosylation profiles differ from bovine/rat IF. Hence, TEV cleavage step, while efficient in removing His‑tags, adds another potential loss point-suboptimal cleavage. Also, it is possible that the precipitation can disproportionately impact already low‑yield species. These wet lab outcomes align with the dry lab observation that bovine IF’s geometry allows tight packing around the ligand (B12), suggesting a well‑folded state that could resist aggregation during purification.
Figure 3. SDS‑PAGE post‑TEV cleavage showing varying yields with rat and bovine IF yielded more prevalent post‑TEV bands; human and platypus IF bands were highly faint, reflecting lower recovery
Achievement 4: SEC Profiling and Selection of Pure Fractions
What we achieved:
Size exclusion chromatography (SEC) separated IF proteins from contaminants and aggregates with fractions selected based on UV peaks and SDS-PAGE confirmation.
Our critical discussion:
SEC peak profiles provided us a clear fingerprint for recombinant IF of each species. Bovine IF’s sharper and symmetrical peak and stronger gel validation indicated higher purity and minimal aggregation compared to others. This is also a trait of bovine that we found with dry lab modelling prediction as linked to its stable fold. Rat IF also showed a pronounced peak, though slightly broader, which could indicate mix of monomeric and oligomeric forms. On the other hand, human IF’s smaller, less defined peak reflects our upstream limitations in yield and perhaps partial degradation. Additionally, the platypus IF peak was less intense, which is consistent with a lower concentration of the sample.
Nevertheless, porcine IF’s absence of a peak and gel band confirmed that upstream expression was insufficient for meaningful recovery for downstream applications. Hence, after SEC profile and band confirmation, our test candidate number was reduced to four: human IF, bovine IF, rat IF and platypus IF.
Figure 4. SEC chromatograms and SDS-PAGE of selected SEC fractions. A) bovine IF's sharp peak indicates higher purity and minimal aggregation, B) Porcine (pig) IF produced no discernible peak in the UV trace and showed no visible bands in SDS‑PAGE
Achievement 5: Functional Validation with B12 Binding Capacity
What we achieved:
We developed a very simple UV‑Vis binding assay to measure cyanocobalamin binding to IF proteins (see Measurement 2), separating bound from free B12 using 3.5 kDa MWCO filters.
Our critical discussion:
Our in-house binding assay showed us functional differences between IF variants. Bovine IF showed high binding efficiency consistent with dry lab docking predictions of strong geometric complementarity. Rat IF was the most unsuccessful one in terms of binding percentage, which is also consistent with its significantly weaker predicted affinity found in dry lab simulations. Platypus IF retained detectable but lower amounts of B12, which may reflect evolutionary divergence in binding site residues despite overall structural integrity. From the human IF side, results revealed that although clinically relevant, it may require expression in a mammalian host to ensure optimal binding with B12.
All in all at this part, we showed that the technical constraints such as low protein concentrations, buffer optimization, and pH control were critical for reproducibility. However, we need to point out the fact that this functional layer confirms that yield alone is not sufficient; so binding assays are essential to identify candidates with both production feasibility and biological activity.
Figure 5. Bar chart of % B12 bound for each species (baseline‑corrected) and standard curve of absorbance at 361nm, linked with dry lab ligand binding overlays
Achievement 6: Immunological Insight: Antibody Binding by ELISA
What we achieved:
We performed indirect ELISA to measure cross‑reactivity of species‑specific recombinant IFs with commercial anti‑human GIF antibodies (see measurement 1).
Our critical discussion:
Although ELISA was not designed to measure vitamin binding, it offered valuable insight into potential antibody interference. Commercial human IF produced the highest absorbance, validating assay performance (positive control). All recombinant IF variants showed markedly lower binding (~33% of commercial IF signal at 1:5,000 primary antibody dilution), with rat IF giving the highest among them. This suggests lower cross‑reactivity for bovine and platypus IF, which could be advantageous in patients with autoimmune gastritis or pernicious anemia where anti‑human IF antibodies block function.
We want to emphasize that rather than being a “negative result,” these low absorbance values can be interpreted as a safety signal as for example non‑human IFs may evade pathogenic antibody binding while retaining B12 binding capacity. At this point, we highlight that incorporating this functional complement extends candidate evaluation beyond yield and affinity.
Figure 6. A) Our ELISA results when substrate added and color change when stop solution is added, B) 96-well plate organization with all control groups and commercial IF as positive control, C) bar chart of dilution‑corrected absorbances (mean ± SD) for each species
Achievement 7: Structural Characterization for our best candidate so far
What we achieved:
We used CD spectroscopy to assess secondary structure content of purified bovine IF, with BSA as positive control (see Measurement 3see ).
Our critical discussion:
CD spectra confirmed that bovine IF adopts a correctly folded α/β architecture, with shallow minima near 208 nm and 222 nm consistent with mixed α‑helical and β‑sheet content, as predicted by BeStSel. The BSA control showed the expected strong α‑helical signature with deep negative bands at 208 and 222 nm. Dry lab modelling predicted bovine IF’s stable, compact fold and high geometric complementarity with B12; the CD data support this by showing a well‑defined secondary structure even at our low protein concentration (~0.074 mg/mL). Overall, it can be said that our structural validation strengthens the case for bovine IF as our lead candidate, bridging computational predictions with biophysical measurement.
Figure 7. Comparison of secondary structure content estimations for Bovine Intrinsic Factor (BOV-IF) and Bovine Serum Albumin (BSA) obtained using A) BeStSel prediction server, B) our in-house Python analysis pipeline. The Python method uses mean residue ellipticity landmarks at characteristic wavelengths (208/222 nm for α-helix, 215 nm for β-sheet, and 200-205 nm for coil) to produce quick, reproducible α/β/coil splits, while BeStSel employs full-spectrum deconvolution optimized for β-rich structures. Results illustrate a mixed α/β fold for BOV-IF and predominantly α-helical character for BSA, consistent with known structural trends
Future Perspectives: How our results and discussions shed light on new-generation oral B12 supplements?
Looking ahead, our data points towards a rational path for developing IF‑based oral B12 supplements. The consistently higher yields, structural stability, and binding efficiency of bovine IF make it a strong production candidate, while rat IF’s high per‑molecule binding affinity suggests it could serve as a template for engineering enhanced variants. By combining these strengths, we hypothesize that a hybrid construct could be designed, for example by using bovine IF backbone for stability with rat IF binding site residues for affinity.
Nevertheless, improving human IF yield remains important for clinical translatability, potentially by co‑expression with folding chaperones or shifting to mammalian or insect cell systems for post-translational modifications that are closer to the native ones [1, 2].
On the other hand, formulation testing in simulated gastrointestinal conditions would assess stability during oral delivery, and binding assays with holo‑transcobalamin could evaluate bioavailability potential in the future experiments. Also, it is very important to emphasize that immunogenicity should be addressed through expanded ELISA testing, especially for non‑human IF variants, to ensure safety in autoimmune contexts.
From structural studies side, including crystallography, which we couldn’t perform due to a lack of time and resources, would allow the mapping of B12 interaction sites at atomic resolution, guiding further engineering. Together, these steps can offer a clear, evidence‑based roadmap from proof‑of‑concept to a viable IF‑based therapeutic as well as being a map for other vitamin supplements, with our cross‑species dataset serving as a foundation for applied product development.
References
- Xiao, S., Shiloach, J., & Betenbaugh, M. J. (2014). Engineering cells to improve protein expression. Current opinion in structural biology, 26, 32–38.
- Gordon, M., Chokshi, H., & Alpers, D. H. (1992). In vitro expression and secretion of functional mammalian intrinsic factor using recombinant baculovirus. Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression, 1132(3), 276–283.
