At Cradle, we need to use every tool in the protein engineering toolbox for high-throughput protein engineering with high data quality and quick turnaround times. Protein tags are one such tool. These molecular "handles"—some just 6 amino acids in size—confer useful properties to our proteins that accelerate our lab workflows and help us validate our machine learning modeled sequences faster and more reliably.

Working with proteins in the lab is not without challenges. Picture your protein of interest as a single puzzle piece in a box of hundreds or thousands of other pieces - hard to find on its own. Protein tags act as small, molecular “handles”–short sequences of amino acids added to the beginning or end of a peptide sequence. They make your protein of interest easier to pick up from the mix - whether you’re ‘fishing’ it out, labeling it, or testing it in a new way.  Over the last 40 years, these versatile tools have completely changed how biologists isolate, detect, quantify, and study proteins.

At Cradle, protein tags (particularly affinity tags) are fundamental to accelerating our protein engineering workflows. Although we're primarily a software company building AI tools for protein design, we maintain an in-house wet lab specifically to validate our ML models and generate high-quality training data. These protein tags accelerate our lab work and enable our researchers to efficiently test hundreds of computationally designed protein variants in just a few days, creating a powerful feedback loop that continuously improves our AI platform. 

The science behind protein tags

At the molecular level, protein tags confer additional properties to engineered proteins, such as high affinity for other molecules or increased solubility. For example, polyhistidine tags (His-tags) add a small extra domain containing multiple histidines that strongly bind to metal ions like nickel. 

The beauty of protein tags lies in their versatility. They can be used for various purposes:

• Purification: Tags like the polyhistidine (His) tag allow us to isolate proteins from a mixture using a method such as affinity chromatography.

• Detection and quantification: Highly-specific epitope tags such as FLAG or HA (Influenza hemagglutinin protein) act as molecular "name tags" that can be detected by specific antibodies in a detection assay.

• Solubility enhancement: Some tags, like the SUMO tag, can help proteins fold correctly or make them more soluble. GST (glutathione S transferase) and MBP (maltose binding protein) confer this advantage as well, in addition to being useful affinity purification tags.

• Visualization: Fluorescent protein tags enable us to track proteins in living cells, revealing their localization and dynamics.

A comparison of common protein tags

Let's compare some of the most commonly used protein tags:

Each tag has strengths and potential drawbacks. The selection of the best tag depends on the specific experimental requirements, protein characteristics, and downstream applications. For instance, if you're working with a protein that tends to form inclusion bodies, using a solubility-enhancing tag like GST, SUMO or MBP might be beneficial. On the other hand, if you need to isolate a high yield of purified protein, the His-tag would be a better choice.

Tag placement considerations

The location of a tag can significantly impact protein function. For example, when working with nanobodies, the N-terminus of the peptide sits very close to the complementarity-determining region (which is critical for binding), making the C-terminus preferable for attaching a tag. Some proteins may tolerate tags at either end, but this must be determined empirically. When onboarding new proteins, testing the effect of tags at either terminus on assay performance is important.

In some cases, internal tags can be used, but these require careful design to avoid disrupting protein structure. Flexible linker sequences (often glycine-serine repeats) can be added between the tag and the protein to minimize interference with folding and function.

Protein tags and their compatibility in high-throughput protein engineering 

At Cradle, we develop AI software for protein engineering. To train and validate our machine learning models, we need to be masters of "high-throughput protein engineering," which means expressing, purifying, and testing hundreds of protein variants in parallel.

In this context, protein tags play a crucial role, and they need to be well suited for automation and high-throughput applications. With the right protein tags, we can dramatically shorten the time required to go from in-silico design to experimental validation.

Protein tags in high-throughput purification

The tags we use in our lab are particularly powerful when combined with automated liquid handling systems and multichannel pipettes. We've optimized protocols for parallel purification of hundreds of protein variants using 96-well magnetic separator blocks and automated plate washers.

Leveraging protein tags effectively in a high-throughput workflow requires selecting both a tag and workflow that suit your needs. For example, at Cradle, we want to rapidly purify high yields of protein in multi-well plates and do not want big tags that will impact protein properties we are optimizing. For this, we use the His-tag to tag our proteins, and commercially-available nickel-coated magnetic beads to purify them.

In other cases, different tags may be better suited. Here's a comparison of tags that we have evaluated for high-throughput purification:

Optimizing high-throughput workflows with protein tags

The efficiency of our high-throughput protein purification depends not only on the tags themselves but on how we strategically integrate their properties to fit our automated workflows. At Cradle, our approach focuses on minimizing complexity while maximizing throughput:

1. Tag selection principles: We deliberately choose purification tags compatible with magnetic beads that work well with our liquid handlers, making our protocols more amenable to automation.

2. Miniaturization-friendly systems: We select tagging systems (like splitGFP) specifically designed to work in high-throughput, plate-based assays, enabling rapid detection and quantification.

3. Streamlined protocols: We prefer small tags that don't require removal steps or interfere with protein function, as tag removal adds time, complexity, and reduces overall yield. 

4. Validation: When onboarding new proteins or tags, they are tested in all of our existing workflows to make sure they have minimal impact on yield, protocols, and assay values.

By systematically selecting tags that meet these requirements and testing the tagged proteins through all our high-throughput protocols and assays in the lab, we've been able to dramatically increase our throughput, purifying and testing hundreds of protein variants in the time it would traditionally take to examine just a handful. This capability is crucial for our machine learning-guided protein design efforts, allowing rapid iteration and validation of in-silico predictions. Let's go a bit deeper into the protein tags that we use and have evaluated for our high-throughput lab.

His-tag: The workhorse

The His-tag, despite being one of the oldest affinity tags, remains a workhorse in high-throughput applications and it is central to Cradle’s purification workflow. Its affinity for nickel ions makes it ideal for use with nickel-coated magnetic beads, enabling rapid purification in 96- or 384-well plate formats. Commercial systems like Ni-NTA magnetic agarose beads are widely available and compatible with automated liquid handling systems, and are the fastest way to get high yields of pure protein. In our lab, we go a step further and utilize the BlueCatBio BlueWasher to accelerate purification of His-tagged proteins even more - see our blog post for details and a full protocol.

Strep-Tag II: Higher specificity

The Strep-Tag II has gained popularity in high-throughput workflows due to its high specificity and mild elution conditions. Strep-Tactin® magnetic beads are commercially available, allowing for highly-specific and rapid parallel purification while preserving protein activity. Using these beads combined with our automated purification workflow, we have achieved high purity and relatively high yield for proteins in our lab, but due to the higher yield and faster purification time of the His-tag, the Strep Tag II remains a back-up.

Other useful tags

Tags like the HaloTag, which forms covalent bonds with its ligand, provide exceptional specificity and are  compatible with stringent washing conditions, resulting in high purity. The HaloTag also has an advantageous short protease that cleaves off the tag after purification in only 90 minutes, much faster than traditional TEV protease cleavage. This tag, while in the end not selected for Cradle’s core workflow due to the increase in time per purification, remains a strong choice for high-yield and extremely highly-specific purifications in other set-ups.

The ever-expanding toolkit of protein tags

While current protein tags have revolutionized how we work with proteins, the field is constantly evolving. Many other strategies and chemistries have been developed that expand a scientist’s toolkit for protein detection, manipulation, and analysis. Some of these include:

1. Click chemistry: Allows for specific, rapid, and efficient labeling of proteins without genetic fusion tags.

2. Intein-mediated purification: Uses self-cleaving protein splicing elements for autocatalyzed removal of tags after protein purification.

3. Tandem Affinity Purification (TAP) tags: Combines two affinity tags for sequential purification steps, resulting in exceptionally pure protein complexes.

4. Proximity labeling: Techniques like BioID, TurboID, and APEX provide tag-based ways to identify protein-protein interactions in living cells without traditional fluorescently-labelled tags.

5. Engineered epitope tags: Improving upon limitations of existing linear epitope tags such as HA and FLAG, engineered tag systems like ALFA-tag and the nanobody to which it strongly binds are increasing the number of applications for which these highly-specific tags can be leveraged.

6. Recombinant protein binding: Tag technologies like the SpyTag/SpyCatcher system add a new tool to the protein tagging toolkit: the ability to recombinantly bind any two proteins together by forming a bond between the tags on each protein.

At Cradle, we're always keeping a close eye on new technologies that could accelerate or increase robustness in our workflows - tags are one of many tools that help us build and test hundreds of proteins at once to generate data for our models. By integrating new technologies in the lab with our machine learning approaches, we are continually developing better and faster ways to engineer proteins.

Further readings

1. Comparison of affinity tags for protein purification

2. Overview of affinity tags for protein purification