When Your Expression System Isn’t Cooperating With Your Protein
Most biologics teams do not begin development by choosing an expression host from a blank slate. They usually start with the platform they know, the host that their organization has already invested in, or the system their CDMO is prepared to run.
That is practical. Established expression systems bring familiarity, existing analytics, internal experience, and a clearer path through development.
But sometimes the target protein does not cooperate with the system being used. When that happens, the molecule can start to be viewed as a “problem molecule.”
Sometimes that label is accurate. But in other cases, the molecule may not be the problem. The problem may be a mismatch between the protein and the expression system being used.
For non-glycosylated biologics, the issue is not simply whether a host can express the protein. The more important question is whether the expression system can produce usable, recoverable, active product without creating avoidable quality, recovery, timeline, or cost burdens.
At Primrose Bio, this is the type of problem Pfenex Expression Technology® is designed to help evaluate: whether a protein’s challenges are primarily construct-related, process-related, or the result of a mismatch between the molecule and the host biology.
That distinction matters because many expression problems do not appear as outright failure. They appear as friction: poor solubility, clipping, aggregation, inconsistent activity, low recovery, difficult purification, or repeated optimization cycles that keep the program moving, but slowly.
Friction is the hidden work required to compensate for poor host – protein fit. It is the gap between an expression result that looks acceptable on a slide and a product stream that actually moves a program forward.
So when these issues appear, the first question should not necessarily be, “Should we look at another expression ?” A better question is, “Is our current expression system working with this molecule, or are we forcing the molecule through a system that is creating friction across development?”
What Host–Protein Mismatch Can Look Like
A host–protein mismatch is not always obvious from the first expression readout. A process may produce measurable protein and still create problems that affect development.
These are the situations in which a promising candidate can be mislabeled as a “problem molecule” and deprioritized before teams know whether the molecule itself is the issue, or whether the expression system is creating avoidable friction.
Common signs include:
- strong total expression but low soluble or active product
- acceptable titer but poor recovery
- clipping or degradation during expression or harvest
- aggregation or inconsistent product quality
- high impurity burden
- repeated purification challenges
- functional activity that does not match the expression result
extensive optimization required before the material is useful for development decisions The first distinction is whether the issue is related to expression design or host biology.
Expression-design issues may involve promoter strength, induction strategy, codon usage, secretion leaders, fusion partners, or vector architecture. Host-biology issues may involve folding environment, protease activity, chaperone availability, localization, metabolic burden, or recovery from the compartment where the protein accumulates.
Both matter. Expression is not just a DNA sequence entering a production system. It is a living organism being asked to make, fold, localize, and protect a molecule it did not evolve to produce.
For non-glycosylated proteins, that distinction matters because the host should be judged less by default platform fit and more by whether it produces functional, recoverable product with a practical path to development.
A familiar platform may still be the right choice, especially when an organization has already invested heavily in the infrastructure, analytics, and know-how required to run it efficiently. But that platform should earn its position based on molecule fit, not simply organizational habit.
Before Changing Expression Systems, Test The Levers That Matter
The first step is not necessarily to abandon the current platform. In many cases, teams can improve performance by testing the levers already available within the system.
For microbial expression, those levers may include promoter selection and design, secretion leader screening, codon usage and vector architecture, fusion partners, periplasmic versus cytoplasmic targeting, protease-deficient strains, folding helpers, chaperone support, media conditions, feed strategy, induction scheme, temperature, and harvest timing.
It may also require looking more closely at which assay is driving the decision. A titer measurement alone may not reveal whether the product is active, intact, properly folded, soluble, or recoverable.
The question is not whether the current platform can be made to work at all. Many systems can be optimized enough to move a program forward. The better question is whether the required optimization is proportionate to the value being created.
Pfenex Expression Technology gives Primrose a way to evaluate multiple expression variables before assuming the candidate protein is inherently the problem. The goal is to identify the lowest-friction path to soluble, intact, recoverable protein that can support development decisions.
A protein that “doesn’t express well” is often not inherently a problem protein. It may be a protein being asked to perform in a mismatched host or expression context.
The Cost Sink That Can Be Easy To Miss
For non-glycosylated proteins, the cost of staying with a familiar system is not always visible as a simple manufacturing number.
The cost may show up as:
- additional strain-engineering or cell line development cycles
- repeated purification or recovery troubleshooting
- delayed analytical characterization
- additional process-development work, whether internal or at a CDMO
- uncertainty about scale-up efficiency
This can be particularly relevant for organizations that rely on CHO expression as their default platform. CHO may remain the right choice when mammalian post translational processing, e.g. glycosylation is needed, or existing platform economics justify it. But for a molecule that does not require glycan biology, the added run time, cost structure, and process complexity of a mammalian system should be actively justified rather than assumed.
In those cases, evaluating a microbial or alternative expression strategy can complement the installed platform by helping teams recognize mismatch early, before the program absorbs unnecessary development burden.
Development Phase: Ample High-Quality Protein to Support Decisions
Titer is useful, but it is not enough.
At the development stage, the better question is whether the system produces decision-grade protein: material that can support meaningful evaluation of binding, activity, stability, purity, recovery, and manufacturability. A high expression number matters less if the product is misfolded, inactive, insoluble, clipped, aggregated, or difficult to recover.
This is why expression-system fit should be evaluated through a broader lens. The right platform should help the molecule move forward, not merely generate a number that looks promising before downstream work begins.
Questions To Ask Before Accepting The Burden
When a molecule is not cooperating with the current expression system, teams should ask:
- Are we measuring total expression, or usable active product?
- Is the molecule failing at expression, folding, localization, recovery, or purification?
- Are quality or recovery problems recurring across constructs or process conditions?
- How much development time is being spent compensating for platform mismatch?
- Does the molecule actually require the host system we are using?
- Would another expression strategy reduce complexity, or introduce more risk than it removes?
These questions do not assume the answer is to change platforms. They help teams decide whether optimization within the current system is enough, or whether the expression strategy itself deserves another look.
A Better Expression-System Conversation
For non-glycosylated biologics, expression system evaluation should not start with a generic host comparison. It should start with molecule fit.
If the current system produces usable, active, recoverable product with a reasonable path to scale, staying the course may make sense. But if the platform requires ongoing workarounds to address folding, recovery, impurity, or quality issues, the true cost of staying with the default may be higher than it appears.
Primrose uses Pfenex Expression Technology to help teams evaluate whether expression challenges are construct-related, process-related, or rooted in host biology — and to identify lower-friction paths to soluble, active, recoverable protein.