Week 1 HW: Final Project Ideas

(1) What you would like to biomanufacture and why?

I want to biomanufacture two recombinant lysosomal enzymes: α-galactosidase A (α-Gal A) for Fabry disease and β-glucocerebrosidase (GCase) for Gaucher disease. Both are currently produced via enzyme replacement therapy (ERT) using mammalian cell culture platforms, which are complex and costly (Hughes, 2025). My motivation comes from my research placement in Professor Derralynn Hughes’ laboratory at the UCL Lysosomal Storage Disorders Unit, where I have been exposed to the clinical realities of these conditions and the limitations of current therapeutic manufacturing.

Fabry disease is an X-linked lysosomal storage disorder caused by reduced or absent α-galactosidase A activity due to variants in the GLA gene, leading to progressive accumulation of globotriaosylceramide (Mehta et al., 2010). Gaucher disease, the most common lysosomal storage disorder, results from mutations in GBA1 encoding β-glucocerebrosidase, causing glucocerebroside accumulation primarily in macrophages and leading to hepatosplenomegaly, haematological abnormalities, and skeletal complications (Grabowski, 2008). Professor Hughes’ research programme focuses on understanding the pathophysiology of phenotypic heterogeneity in Fabry disease and bone-related pathology in Gaucher disease, and she has authored over 150 papers in the area of macrophage biology and lysosomal storage disorders (Hughes, 2025).

The clinical burden of these diseases is significant. Fabry disease has an estimated global prevalence ranging from 1 in 40,000 to 1 in 170,000 individuals (Mehta et al., 2010). Gaucher disease affects approximately 1 in 40,000 to 60,000 individuals worldwide, with higher prevalence in Ashkenazi Jewish populations (Grabowski, 2008). Despite the availability of ERT, treatment costs exceed hundreds of thousands of dollars annually per patient, creating substantial barriers to access in underfunded healthcare systems (Hughes et al., 2025). Gaucher disease in particular represents a significant clinical burden, and GBA1 mutations have been identified as major risk factors for Parkinson’s disease, suggesting that therapeutic advances in this area could have implications beyond the LSD field itself (Grabowski, 2008).

My specific target output is purified, active recombinant human α-Gal A and GCase produced in a yeast expression system, intended for preclinical evaluation and eventual formulation as ERT biologics. The underlying engineering challenge is not merely producing the enzyme but producing it with the correct post-translational modifications that enable cellular uptake via the mannose-6-phosphate receptor pathway, which is essential for therapeutic efficacy (Calhoun et al., 2000).

(2) What living system would you use?

I propose using the methylotrophic yeast Pichia pastoris (Komagataella phaffii) as the production chassis, specifically the GlycoSwitch® platform for GCase production, which has been engineered to produce glycoproteins with modified N-glycan profiles that more closely resemble mammalian glycans (Bayraktar et al., 2026).

Pichia pastoris offers several advantages for this application. It achieves high cell-density fermentation in minimal media, secretes recombinant proteins into the culture supernatant at levels substantially higher than mammalian cell systems, and can be genetically manipulated with relative ease (Calhoun et al., 2000). A foundational study demonstrated that recombinant human α-galactosidase A produced in Pichia pastoris was secreted at levels more than 30-fold greater than previously used methods, and the purified enzyme was taken up by fibroblasts derived from Fabry disease patients, restoring normal enzyme levels (Calhoun et al., 2000). This establishes a direct proof-of-concept for the therapeutic potential of yeast-produced α-Gal A.

More recently, human β-glucocerebrosidase has been produced using the Pichia pastoris GlycoSwitch® platform, with both wild-type and mutant (Glu235Ser) GCase expressed as functional glycoproteins. The mutant exhibited higher catalytic efficiency, with approximately 20-fold and 10-fold increases in kcat/KM toward pNPG and 4-MUG substrates, respectively, compared to wild-type (Bayraktar et al., 2026). Notably, in PMA-differentiated U937 macrophage-like cells, which are directly relevant to Gaucher disease pathology, given that Gaucher cells are lipid-laden macrophages, the mutant showed increased cellular uptake and prolonged intracellular activity compared to wild-type (Bayraktar et al., 2026).

At a high level, the manufacturing process works as follows:

• Inputs: Methanol or glycerol/methanol mixtures as carbon source, yeast nitrogen base, biotin, and a buffered medium for optimal secretion. Codon-optimised cDNA encoding human α-Gal A or GCase, cloned into a secretion vector with the α-factor signal sequence (Calhoun et al., 2000).

• Biological process: The yeast cells grow to high cell density under fed-batch conditions, and the heterologous enzyme is secreted into the culture supernatant under the control of the AOX1 promoter. The secreted enzyme undergoes N-linked glycosylation in the yeast endoplasmic reticulum and Golgi apparatus (Calhoun et al., 2000; Bayraktar et al., 2026).

• Output: After fermentation, the secreted enzyme is purified from the clarified supernatant using ion exchange and hydrophobic interaction chromatography, as demonstrated in the original α-Gal A work (Calhoun et al., 2000).

Assumptions: I assume the GlycoSwitch® platform can be adapted for co-expression or parallel production of both enzymes in the same facility. I also assume that the N-glycan profiles produced by this platform are sufficient for mannose-6-phosphate receptor-mediated uptake, though this would require experimental verification (Bayraktar et al., 2026).

(3) How would you use automation?

I would use a two-tiered automation strategy: an Opentrons OT-2 liquid-handling robot for high-throughput strain screening and process optimisation at laboratory scale, combined with the MIT Cloud Lab for Learning and its Reconfigurable Automation Carts (RACs) for larger-scale fermentation monitoring and remote programming (Petzold & Mukhopadhyay, 2025).

The most impactful application of automation in this workflow is clone screening and expression optimisation. The APEX pipeline has demonstrated automated recombinant protein production in E. coli using the accessible, open-source Opentrons OT-2 platform with high precision and reproducibility (Petzold & Mukhopadhyay, 2025). An analogous workflow for Pichia pastoris would involve automated transformation, colony picking, microcultivation in 96-well plates, and automated sampling for enzyme activity assays. This is particularly important because expression levels in Pichia pastoris are highly dependent on cDNA copy number and culture conditions. The original α-Gal A study demonstrated that expression increased with increasing cDNA copy number at 25°C, but not at the standard growth temperature of 30°C; this would have required extensive, labour-intensive screening to discover manually (Calhoun et al., 2000).

The OT-2 platform can also automate the enzyme activity assays critical for evaluating production batches. For GCase, the activity can be measured using fluorogenic substrates such as 4-methylumbelliferyl-β-D-glucopyranoside (4-MUG) and p-nitrophenyl-β-D-glucopyranoside (pNPG), which are amenable to microplate-based assays (Bayraktar et al., 2026). For α-Gal A, similar fluorogenic substrates are available. Automated sampling and assay assembly would enable real-time monitoring of enzyme production and specific activity throughout the fermentation process.

For larger-scale production, I would use the MIT Cloud Lab to program and monitor bioreactor runs remotely (Petzold & Mukhopadhyay, 2025). The Cloud Lab’s 6 Reconfigurable Automation Carts would allow automated sampling for enzyme activity and glycosylation analysis, real-time adjustment of methanol feed rates, and remote monitoring of dissolved oxygen and pH. This approach aligns with the growing emphasis on robotics and automation in biomanufacturing, as highlighted by the course’s focus on cloud labs, desktop liquid-handling robots, and programmable microfluidics (Asin-Garcia et al., 2025).

The specific workflow I would automate first is parallel micro-scale expression screening of different GCase and α-Gal A constructs: varying cDNA copy number, signal sequences, and cultivation temperatures across 96-well plates, with automated sampling and enzyme activity reads. This would generate a dataset linking construct design to expression level and specific activity, directly informing the choice of production strain for scale-up (Petzold & Mukhopadhyay, 2025).

(4) How much would you need to make, and how long would it take?

(a) How much?

I will define a meaningful scale as producing enough enzyme to supply a one-year treatment course for 100 patients with Fabry disease or Gaucher disease.

For Fabry disease, the recommended dose of agalsidase beta (Fabrazyme) is 1 mg/kg body weight every two weeks. For a 70 kg adult, this is 70 mg per infusion, or approximately 1.82 g per patient per year (Mehta et al., 2010). For 100 patients, the total annual requirement would be 182 g of purified α-Gal A.

For Gaucher disease, the recommended dose of imiglucerase (Cerezyme) is 2.5 mg/kg body weight every two weeks for the initial treatment phase (Grabowski, 2008). For a 70 kg adult, this is 175 mg per infusion, or approximately 4.55 g per patient per year. For 100 patients, the total annual requirement would be 455 g of purified GCase. However, for a pilot-scale demonstration, I will focus on producing 10 g of purified GCase, which represents approximately 2% of the annual requirement for 100 patients and is sufficient for preclinical evaluation and process validation.

(b) How long?

Using published performance data for Pichia pastoris high-cell-density fermentation:

α-Gal A production: The original study reported that the level of production was more than 30-fold greater than that of previously used methods (Calhoun et al., 2000). While specific productivity data in g/L are not provided in the abstract, typical Pichia pastoris secretion levels for heterologous proteins range from 0.1 to 1 g/L in optimized high-cell-density fermentations. I will use a conservative estimate of 0.5 g/L of secreted α-Gal A in the culture supernatant.

GCase production: The GlycoSwitch® study produced functional GCase but did not report volumetric productivity in the abstract (Bayraktar et al., 2026). Based on published Pichia pastoris fermentation data for other secreted enzymes, I will estimate a productivity of 0.2 g/L for GCase, assuming that the GlycoSwitch® glycosylation engineering imposes a modest metabolic burden (Kazemi Seresht et al., 2013).

Working volume: I will assume a 10 L working volume in a pilot-scale bioreactor for initial production, with the potential to scale to 100 L.

Calculation for GCase (10 g target):

GCase production per batch = 0.2 g/L × 10 L = 2 g

Number of batches required = 10 g ÷ 2 g/batch = 5 batches

Fermentation time per batch: Seed train (24 h) + batch growth phase (24 h) + methanol induction phase (72–96 h) = 5–6 days per batch

Total fermentation time = 5 batches × 5.5 days = approximately 28 days

With parallel operation of two bioreactors, the total production time could be reduced to approximately 14 days.

Assumptions and caveats:

• I am assuming that the productivity of 0.2 g/L at 10 L scale can be maintained. Scale-up from laboratory scale to pilot scale often results in some loss of productivity due to reduced oxygen transfer and increased shear stress. A scale-up factor of 50–70% of lab-scale productivity would increase the required number of batches to 7–10 (Kazemi Seresht et al., 2013).

• Purification yields are not included. Typical recovery yields for ion exchange chromatography are 60–80%, so starting with 20 g of secreted GCase would yield approximately 12–16 g of purified enzyme (Calhoun et al., 2000).

• The metabolic burden of protein secretion in yeast is a well-documented limitation. High-level production of heterologous proteins imposes a burden on host cell metabolism, affecting cellular physiology and potentially limiting specific productivity (Kazemi Seresht et al., 2013). This would need to be characterized experimentally.

5. What don’t you know yet?

Three major unknowns remain:

1. Glycosylation fidelity and mannose-6-phosphate receptor targeting. The therapeutic efficacy of ERT for Fabry and Gaucher disease depends critically on the enzyme’s ability to be taken up by target cells via the mannose-6-phosphate (M6P) receptor pathway. While the original α-Gal A study demonstrated uptake by Fabry patient fibroblasts, the glycosylation analysis indicated the predominant presence of N-linked high-mannose structures rather than complex carbohydrates (Calhoun et al., 2000). High-mannose structures can be taken up by mannose receptors on macrophages—which is relevant for Gaucher disease—but may not be optimal for targeting other cell types. I would need to characterize the N-glycan profiles of both enzymes produced in the GlycoSwitch® platform and measure M6P content and receptor-mediated uptake efficiency in relevant cell models (Bayraktar et al., 2026).

2. Metabolic burden and secretion bottleneck. Protein secretion in yeast is generally associated with a burden to cellular metabolism, and overcoming this burden is a significant engineering challenge (Kazemi Seresht et al., 2013). The secretion bottleneck could be caused by the combination of multi-copy plasmids and the regulation of expression, and may ultimately limit the ability to sustain specific protein secretion. For co-production of two enzymes or high-level production of a single enzyme, I would need to measure the metabolic flux redistribution and identify the limiting steps—whether at the level of transcription, translation, ER translocation, or Golgi processing (Kazemi Seresht et al., 2013).

3. Scale-up of fed-batch fermentation with methanol induction. Methanol feeding strategies are critical for Pichia pastoris high-cell-density fermentation. The methanol feeding rate must be carefully controlled to maintain induction while avoiding methanol accumulation to toxic levels (Calhoun et al., 2000). Scale-up from 10 L to 100 L introduces challenges in mixing, oxygen transfer, and heat removal that could affect both productivity and product quality. I would need to characterize the kLa of the pilot bioreactor and develop a methanol feeding strategy that maintains consistent specific productivity across scales (Petzold & Mukhopadhyay, 2025).

AI Use Disclosure

I used DeepSeek to assist with literature research and structural planning. The specific prompts I used were:

“Find recent papers on recombinant α-galactosidase A and β-glucocerebrosidase production in Pichia pastoris.”

“Search for Opentrons OT-2 automation protocols for microbial strain construction and enzyme activity assays.”

“Provide data on global prevalence of Fabry and Gaucher disease and current ERT costs.”

“Summarize the GlycoSwitch® platform for recombinant glycoprotein production in Pichia pastoris.”

All calculations, assumptions, and interpretations in this assignment are my own. I have verified the key quantitative data against the cited primary literature and will present my understanding of the biomanufacturing workflow in recitation. In particular, the connection between the GlycoSwitch® glycosylation profile and mannose-6-phosphate receptor targeting is an original synthesis that I developed from reading the primary literature on both enzyme production and ERT mechanism of action.

References

Asin-Garcia, E., Fawcett, J. D., Batianis, C., & Martins dos Santos, V. A. P. (2025). A snapshot of biomanufacturing and the need for enabling research infrastructure. Trends in Biotechnology, 43(5), 1000–1014. https://doi.org/10.1016/j.tibtech.2024.10.014

Averesch, N. J. H., Berliner, A. J., Nangle, S. N., Zezulka, S., Vengerova, G. L., Ho, D., Casale, C. A., Lehner, B. A. E., Snyder, J. E., Clark, K. B., Dartnell, L. R., Criddle, C. S., & Arkin, A. P. (2023). Microbial biomanufacturing for space-exploration—what to take and when to make. Nature Communications, 14, 2311. https://doi.org/10.1038/s41467-023-37910-1

Bayraktar, S., Aysan, A., Turunen, O., Valjakka, J., & Binay, B. (2026). Effect of Glu235Ser on enzyme performance and in vitro characterization of human β-glucocerebrosidase in a Pichia pastoris. International Journal of Biological Macromolecules, 361, 151928. https://doi.org/10.1016/j.ijbiomac.2026.151928

Chen, Y., Jin, M., Egborge, T., Coppola, G., Andre, J., & Calhoun, D. H. (2000). Expression and characterization of glycosylated and catalytically active recombinant human α-galactosidase A produced in Pichia pastoris. Protein Expression and Purification, 20(3), 472–484. https://doi.org/10.1006/prep.2000.1325

Grabowski, G. A. (2008). Phenotype, diagnosis, and treatment of Gaucher’s disease. The Lancet, 372(9645), 1263–1271. https://doi.org/10.1016/S0140-6736(08)61522-6

Hughes, D. A. (2025). Lysosomal storage disorders. In Wiley Online Books. Royal Free London NHS Foundation Trust and University College London.

Karim, A. S., & Jewett, M. C. (2021). Toward sustainable, cell-free biomanufacturing. Current Opinion in Biotechnology, 69, 136–144. https://doi.org/10.1016/j.copbio.2020.12.012

Kazemi Seresht, A., Cruz, A. L., de Hulster, E., Heijnen, J. J., & van Gulik, W. M. (2013). Overcoming the metabolic burden of protein secretion in Schizosaccharomyces pombe—A quantitative approach using 13C-based metabolic flux analysis. Metabolic Engineering, 20, 1–10. https://doi.org/10.1016/j.ymben.2013.11.002

Mehta, A., Beck, M., Eyskens, F., Feliciani, C., Kantola, I., Ramaswami, U., Rolfs, A., Rivera, A., Waldek, S., & Germain, D. P. (2010). Fabry disease: A review of current management strategies. QJM: An International Journal of Medicine, 103(9), 641–659. https://doi.org/10.1093/qjmed/hcq117

Petzold, C. J., & Mukhopadhyay, A. (2025). From bench to biofactory: High-throughput technologies and automated workflows to accelerate biomanufacturing. Current Opinion in Biotechnology, 94, 103320. https://doi.org/10.1016/j.copbio.2025.103320

Yap, W. S., Choudhury, D., & Suntornmond, R. (2023). Towards biomanufacturing of cultured meat. Trends in Biotechnology, 41(3), 292–294. https://doi.org/10.1016/j.tibtech.2023.01.014

Ye, J.-W., Lin, Y.-N., Yi, X.-Q., Yu, Z.-X., Liu, X., & Chen, G.-Q. (2023). Synthetic biology of extremophiles: A new wave of biomanufacturing. Trends in Biotechnology, 41(3), 342–357. https://doi.org/10.1016/j.tibtech.2022.11.010

Zhang, Y.-H. P., Sun, J., & Ma, Y. (2017). Biomanufacturing: History and perspective. Journal of Industrial Microbiology & Biotechnology, 44, 773–784. https://doi.org/10.1007/s10295-016-1863-2