Advanced Yeast Glycoengineering

Advanced Yeast Glycoengineering (Pichia) for Biobetters

Microbial systems, human glycans—delivered with composure. A specialised CDMO programme that turns Komagataella phaffii (Pichia) into a dependable factory for custom glycoforms: afucosylated for ADCC, bisected for potency, α2,6-sialylated for PK and anti-inflammatory tone. From strain design to Grade A isolator fill–finish, we make the path feel straight.

Mika Biologics, Microbial CDMO

Executive overview

Most “human-like” glycans in yeast still behave like approximations on inspection. The answer is not a single knockout or a clever tag; it is a disciplined stack: chassis selection, pathway rewiring, secretion architecture, media and redox orchestration, and an analytical regime that sees what matters before it drifts. Mika Biologics builds that stack.

We stand up menus of fucosylation, galactosylation, sialylation (α2,3 and α2,6), and bisecting GlcNAc (MGAT3) with a process that survives scale. We fingerprint each lot by LC–MS at both the released-glycan and glycopeptide levels, lock comparability bands early, and link the glycans you ordered to the behaviour you promised—ADCC, CDC, receptor engagement, and pharmacokinetics that read clean in review.

If you have a yeast/fungal expression plan already, this service interlocks with our Yeast & Fungal Expression Systems page. For method depth and sample COAs, see Analytical & QC for Microbial Biologics. For isolator practice and lyophilisation, see Formulation & Aseptic Fill-Finish (Grade A Isolators). Lifecycle control is anchored in our Process Characterisation, PPQ & CPV for Microbial Platforms.

What we deliver

  • Custom glycoform panels targeted to function:
    • Afucosylated (FUT8-) for FcγRIIIa-driven ADCC.
    • Low-fucose (tuned rather than absolute) when PK and manufacturability prefer a compromise.
    • α2,6-sialylated (ST6GAL1) for anti-inflammatory bias and extended half-life; α2,3 variants (ST3GAL family) where biology requires it.
    • Bisecting GlcNAc (MGAT3) to modulate Fc geometry, often elevating ADCC without pushing fucose to zero.
    • Defined galactose states: G0/G0F, G1/G1F, G2/G2F with tight % bands by site.
    • Man5/Man5-rich for receptor-mediated uptake or ADC precursor work.
  • Chassis and pathway engineering in Pichia:
    • Mannose-extension control (e.g., OCH1 and allied α-1,6/α-1,2 mannosyltransferase logic) to prevent hypermannosylation.
    • GnTI/GnTII (MGAT1/MGAT2) modules to introduce complex-type branching; B4GALT1 for terminal galactose; FUT8 for core fucose (or its removal); ST6GAL1/ST3GAL for sialic acid capping; MGAT3 for bisecting GlcNAc.
    • CMP-Sia pathway support (CMAS, NANS/NANP) and SLC35A1 transport logic for Golgi import where required.
  • Expression architecture optimised for secretion: α-factor pre-pro signal or rational alternatives; AOX1 or GAP promoter strategy (methanol-free where the programme benefits), Kex2 processing harmonised with folding catalysts (PDI, Ero1).
  • Scale that behaves: 5–30 L development; 200–2,000 L GMP-ready; routes to 10,000 L+ through partners without losing the glycan envelope.
  • Analytical fingerprints you can defend:
    • Released N-glycan HILIC-UPLC-FLR with ESI-MS; linkage-specific sialic acid derivatisation when needed.
    • Glycopeptide LC–MS/MS for site-specific occupancy and microheterogeneity.
    • Exoglycosidase maps; DMB sialic acid quantification; fucose %, galactose %, sialylation %, bisecting % as sentinel numbers.
  • Comparability protocols aligned to ICH Q5E/Q6B: pre-agreed acceptance bands for glycan distributions, site occupancy, potency, and higher-order structure; statistical rules that prevent over-reacting to noise or under-reacting to shift.

Use-cases we target

  • Afucosylated antibodies for oncology and infectious disease—maximised ADCC without unstable processes.
  • Bisected Fc fusions where immune engagement is tuned rather than blunted.
  • Sialylated anti-inflammatory biologics (α2,6), including Fc-fusions and cytokine-Fc constructs seeking improved PK/PD.
  • Enzymes and scaffolds that need human-type glycans for stability or receptor traffic but prefer microbial simplicity over mammalian overhead.

The glycoform menus

  1. Core fucose
    • Afucose target: ≤1.0% fucosylated species at the Fc sites; site-resolved confirmation by glycopeptide MS.
    • Low fucose: 5–15% when PK or immunogenicity modelling dictates moderation.
    • Tuning levers: FUT8 expression attenuation; GDP-fucose supply control; culture copper/redox policy (when transferase co-factors are in play).
  2. Galactose state
    • G0/G0F, G1/G1F, G2/G2F with ±5–10% bands per site; B4GALT1 expression and UDP-galactose transport balanced to avoid over-galactosylation.
  3. Sialylation
    • α2,6 preference for anti-inflammatory biology; α2,3 where the receptor landscape requires it.
    • Total sialic acid: 5–25 mol% as designed; linkage split quantified by linkage-specific derivatisation or diagnostic fragments.
    • CMP-Sia flux stabilised by CMAS expression and Golgi import; avoid Neu5Gc contamination by design.
  4. Bisecting GlcNAc
    • MGAT3 occupancy at Fc sites targeted to 10–40% depending on ADCC/PK aims; tracked by site-specific glycopeptide MS.
    • Bisecting plus afucose combinations validated for manufacturability—no “hero lots.”
  5. High-mannose window
    • Where desired for uptake or as a process intermediate, Man5–Man9 % bands are written and defended.

Programme design: QTPP → CQAs → CPPs

QTPP (examples)
Route and dose (IV, SC; mg/kg), target glycan state (afucose ≤1%, α2,6-Sia ≥10%, MGAT3 ≥15%), storage (2–8 °C or lyo), container (type I vial/PFS), shelf life (12–24 months).

CQAs (glyco-centric)
Glycan profile (released and site-specific), fucose %, galactose %, sialylation % with linkage split, bisecting %, high-mannose %, occupancy per glycosite, charge variants (icIEF), aggregates (SEC-MALS), potency (ADCC reporter, CDC, receptor binding), HCP/DNA, β-glucan/yeast carbohydrate carryover, proteolysis markers.

CPPs (that actually move CQAs)
Dissolved oxygen and methanol (if AOX1), pH and redox set-points, copper/manganese where transferases are co-factor sensitive, feed composition (UDP-sugar precursors), temperature ramps, induction timing, residence time in secretory pathway (governed by expression rate and folding assistance), TFF shear and residence, polishing column load and conductivity windows.

Strain and pathway engineering

  • Chassis: glyco-calm Pichia lines with OCH1 deletion or equivalent early-mannosylation control as the base; engineered to prevent outer-chain mannose extension.
  • Humanisation modules: GnTI/GnTII (MGAT1/2) to build biantennary cores; B4GALT1 for terminal galactose; FUT8 for core fucose (add-back or set-down).
  • Sialylation: NANS/NANP to form sialic acid, CMAS to activate it (CMP-Sia), SLC35A1 to import into Golgi, ST6GAL1 or ST3GAL4/6 to cap; tuned to avoid over-capping and preserve charge variants within spec.
  • Bisecting: MGAT3 expressed under a promoter with predictable copy-number behaviour; matches expression rate to Golgi dwell time so bisecting occupancy is stable.
  • O-glycan hygiene: selective PMT family modulation to keep O-mannose under thresholds without compromising viability; verify by glycopeptide scans on serine/threonine-rich motifs.
  • Secretion: α-factor signal variants screened; Kex2 and Ste13 processing alignment so the N-terminus is correct and protease clipping is absent; PDI/Ero1 assistance for disulphide-rich Fc scaffolds.

Upstream (USP) that preserves glyco intent

  • Promoter strategy: GAP for steady methanol-free expression when regulators or HSE policies prefer it; AOX1 where productivity and folding benefit—managed with calm, narrow methanol windows.
  • Bioreactors: single-use stirred tanks (2–2,000 L); low-shear impellers, oxygen transfer sized to promoter; redox controlled to stable glutathione ratio.
  • Feeds: chemically defined, no peptone noise; sugar-nucleotide precursor balance set so glycan occupancy and branching hold in scale-up.
  • IPC: biomass, specific productivity, methanol (if used), dissolved oxygen, pH, base consumption, protease activity, and at-line glycan sentinels (rapid released-glycan HILIC scans in development lots to course-correct early).

Downstream (DSP) with glycan integrity in mind

  • Capture: Protein A for Fc-bearing constructs; affinity/IMAC for tagged development proteins; ion exchange for enzymes and Fc-less scaffolds.
  • Intermediate polish: IEX/HIC trains that avoid harsh conditions known to destabilise terminal sialic acids or trim galactose; conductivity and pH ramps profiled for glycan safety.
  • UF/DF: staged concentration with conservative shear; diafiltration into formulation that prevents desialylation and carbonate-driven pH drift.
  • Host-derived carbohydrate control: β-glucan/mannan removal steps validated; negative marker assays (TLR2/4 signalling proxies where justified) used during development to confirm cleanliness.
  • Protease hygiene: fermentor protease signatures mapped; polishing adjusted or protease clearance step added if clip motifs appear in LC–MS peptide maps.

Formulation and aseptic fill–finish

  • Buffers: histidine or citrate/phosphate for Fc constructs; ionic strength tuned to suppress self-association without stripping terminal sialic acids.
  • Excipients: sugars/polyols and amino acids to support glass transition for lyo; surfactants judiciously—set by container-closure interactions, not habit.
  • Sialic acid preservation: anti-oxidant policies where appropriate; control of trace metals that catalyse loss during storage.
  • Presentations: vials, prefilled syringes, cartridges in Grade A isolators (Grade B background); CCIT deterministic methods set from the beginning; glide-force studies for PFS.
  • Storage: 2–8 °C liquid or lyo at room temperature depending on route and shipping reality; thaw/refreeze limits written into the label.

(Execution detail sits with Formulation & Aseptic Fill-Finish (Grade A Isolators).)

Analytics that make reviewers relax

Released-glycan profile

  • HILIC-UPLC-FLR with ESI-MS for composition and relative abundance; 2-AB or 2-AA labels; exoglycosidase confirmations when needed.
  • Linkage logic for sialylation (α2,3 vs α2,6) using derivatisation fingerprints or MS/MS diagnostics.
  • Fucose % and bisecting % reported as primary sentinels.

Glycopeptide, site-specific

  • LC–MS/MS on tryptic or Lys-C peptides; quantification per glycosite; occupancy verified; site heterogeneity trends used for comparability.

Higher-order and function

  • SEC-MALS, icIEF, DSC/nanoDSF, HDX-MS (where indicated), SPR/BLI for FcγRIIIa and FcRn.
  • ADCC reporter, CDC, and receptor-binding potency assays tied to the intended mechanism.
  • PK surrogates (FcRn binding, sialylation correlation) and immunogenicity flags (non-human epitopes absent by design).

Safety and cleanliness

  • HCP (yeast), residual DNA, β-glucan/mannan, bioburden/sterility; extractables/leachables policy aligned to the presentation.

Fingerprints that travel

  • Each batch produces a Glycan Fingerprint (released + glycopeptide) used for comparability across scale/site/cycle; stored in our LIMS and surfaced on your CPV dashboard.
Mika Biologcs CDMO logo and slogan

Stability to ICH, focused on glycans

  • Long-term (2–8 °C) and accelerated arms; stress (freeze–thaw, agitation, light).
  • Stability-indicating methods include glycan drift (loss of sialic acid, galactose trimming), aggregates, charge variants, potency.
  • Reference standards and trending prevent silent creep; acceptance criteria reflect biology (e.g., ADCC floors tied to afucose bands).

Comparability, PPQ and CPV

  • Comparability protocols lock before you scale: which glycan bins matter, what bands apply, what statistical rules govern pass/fail. We follow ICH Q5E/Q6B intent without turning the exercise into a ritual.
  • PPQ with a qualified scale-down model: three consecutive batches, CPP envelopes proven, CQA variability within agreed tolerance.
  • CPV dashboards track historian tags (feed rates, DO/pH trends, methanol if present, conductivity profiles, column ΔP) against glycan sentinel shifts; you can see movement before QA has to speak.

(Our Process Characterisation, PPQ & CPV for Microbial Platforms page sets out the instrumentation and data model.)

Illustrative project plans

Plan A — Afucosylated Fc therapeutic (ADCC-maximised)

  1. Feasibility (6–10 weeks): Chassis evaluation, FUT8 strategy, early HILIC fingerprints, ADCC reporter proof.
  2. Development (10–16 weeks): DoE on feeds and redox; site-specific glycopeptide MS locked; polishing tuned to protect glycans; ADCC potencies tied to afucose bands.
  3. GMP campaign (8–12 weeks): Three DS lots; DP fills in isolators; PPQ protocol issued; comparability gate ready for inspectors.

Plan B — α2,6-sialylated Fc-fusion (PK & anti-inflammatory)

  1. Feasibility: CMP-Sia pathway and ST6GAL1 tuning; linkage-specific sialylation proof; FcRn binding set.
  2. Development: Formulation to stabilise sialylation; lyo cycle if justified; charge variant control.
  3. GMP: DS/DP, stability start, comparability with clear α2,6/α2,3 linkage bands.

Plan C — Bisected + low-fucose antibody (balanced ADCC/PK)

  1. Feasibility: MGAT3 expression strategy; bisecting occupancy measurement at Fc sites; functional read-through.
  2. Development: Occupancy stability across scale; DoE on copper/redox for transferases; SEC-MALS/ADCC ties.
  3. GMP: PPQ with bisecting % in band; CPV hooks installed.

Why Pichia for biobetters—when it is engineered properly

  • Speed and cost: microbial growth rates and straightforward fermentation beat mammalian timelines for many modalities.
  • Secretion and simplicity: well-behaved secretory pathways when tuned; uncomplicated media; resilient to variable raw-material markets.
  • Glyco-agency: unlike a “black-box” mammalian default, you can set glycoform states as intentional design variables—and keep them there.

Inter-page guidance

  • Pair with Yeast & Fungal Expression Systems if you are still choosing a host.
  • Use Analytical & QC for Microbial Biologics for method transfer SLAs and example COAs.
  • For presentations and logistics, see Formulation & Aseptic Fill-Finish (Grade A Isolators).
  • For lifecycle documents and dashboards, Process Characterisation, PPQ & CPV for Microbial Platforms keeps the thread unbroken.

FAQ

Can Pichia truly deliver human-like complex-type N-glycans?
Yes—when early mannosylation is constrained and GnTI/GnTII, B4GALT1, FUT8, and the sialylation rail are integrated with transport and flux control. We prove it at the released-glycan and site-specific levels.

How do you choose α2,6 vs α2,3 sialylation?
We decide with you, based on biology. α2,6 is often preferred for anti-inflammatory tone and PK; α2,3 for certain receptor landscapes. We install the appropriate sialyltransferase and confirm linkage by diagnostic LC–MS chemistry.

Will afucose always maximise ADCC?
Afucose lowers FcγRIIIa KD and typically elevates ADCC. Some programmes prefer low-fucose plus bisecting for manufacturability or PK. We test both and write acceptance bands around performance, not fashion.

How is hypermannosylation prevented?
Through OCH1-lineage control and careful management of α-mannosyltransferases, plus process conditions that avoid stress-induced rescue pathways. The proof is in the LC–MS spectrum, not a single genotype.

What does an LC–MS glycan fingerprint look like?
Released-glycan HILIC chromatograms with annotated peaks and ESI-MS confirmation; a matching glycopeptide table per site. We maintain a fingerprint library for comparability and CPV trends.

Can you execute comparability if we move from CHO to Pichia?
Yes—with clear function mapping. We define the new glycan spec, demonstrate mechanism continuity (e.g., ADCC, receptor binding), and present a statistical comparability argument aligned to ICH Q5E/Q6B.

Do you support bisecting GlcNAc?
We do. MGAT3 is integrated and tuned; occupancy is measured per Fc site. We provide combinations with afucose or low-fucose as required.

What about O-glycans and yeast carbohydrate residues?
We measure them directly (glycopeptide scans, β-glucan/mannan assays) and set limits. Processes are validated to keep them silent in safety read-outs.

What yields can we expect?
Programme-dependent. Our design emphasises predictable secretion and stable glycoform distribution over record titres that collapse during PPQ. We share titre corridors after feasibility.

Are there methanol-free options?
Yes. GAP-driven designs are common for regulatory or operational reasons. AOX1 remains available where it wins on productivity—with a safety case and narrow control.