https://orcid.org/0000-0001-8760-6731
Proteomics is a powerful tool for studying disease, yet massively underused in clinical laboratory settings. To achieve wide utilization across clinical laboratories, proteomics technologies must become cheaper and faster, and require less expertise to operate. In response to this recognized need, the last 5 years have seen rapid advances towards the potential use of nanopores for sequencing proteins. In recent work published in Nature Methods, Wang and colleagues (1), associated with the Huang lab at Nanjing University, demonstrated a major new milestone in this field.
Nanopore sequencing for RNA and DNA has had a profound impact on scientific research in human genomics. Long-read, single-molecule sequencing, coupled with a small physical footprint, low cost of operation, and real-time data generation, has the potential for far-reaching clinical applications. These sequencing methods use a motor protein to pass polymers across an electrically resistant membrane through a transmembrane protein pore. The pore has been modified to change the electrical resistance of the local membrane, producing a measurable change in ionic current that can be associated with each passing nucleic acid. In concept, this approach can be applied to a wide variety of biopolymers, including proteins.
One approach to nanopore sequencing of proteins utilizes peptides as intact polymers that can be fed into a nanopore using similar constructs as those developed for nucleic acid sequencing. For example, Brinkerhoff et al. (2) developed a nanopore construct with multiple motor proteins that ratcheted a single peptide polymer into the nanopore at a fixed rate using a DNA tether. At present, amino acid-level resolution is not possible as the membrane current is affected by nearby amino acids. Furthermore, amino acid side chains (R groups) are significantly smaller than nucleic acids and thus have a more limited chemical space; thus, discerning modified amino acid residues is difficult, let alone resolving the hundreds of protein posttranslational modifications (PTMs) which may occur.
While R group recognition in an intact polymer remains challenging, single amino acids contain both amino and carboxylic acid functional groups that can provide a broader scope for chemical interactions. As such, an alternative approach to nanopore sequencing relies on hydrolyzing a trapped peptide or protein using a proteosome-coupled nanopore (3) and measuring the resulting amino acids, as shown in Fig. 1A. While contextual sequence-ordering data is lost when digesting peptides down to amino acids, the relative abundances of amino acids in a peptide can act as an identifiable fingerprint. Previously, Ouldali et al. (4) demonstrated that it was possible to recognize amino acids using an aerolysin-based nanopore, but they could only differentiate 13 of 20 proteinogenic amino acids. The Huang lab demonstrated that an alternative nanopore, Mycobacterium smegmatis porin A (MspA), could also be engineered to recognize amino acids (5). MspA forms an octameric protein nanopore that has been popular for DNA sequencing due to its narrow constriction site. In this new work (1), the Huang lab showed that by changing a single aspartic acid (N90) in one MspA subunit to a cysteine modified with a nickel-nitrilotriacetic acid (Ni-NTA), the hetero-octameric MspA mutant could recognize all 20 proteinogenic amino acids with 98.8% accuracy. The authors also demonstrated the accurate detection of 4 classes of biologically important PTMs: dimethylarginine, acetylated threonine, N-GlcNAcylated-aspargine, and phosphoserine. That said, it is important to note that while this approach can identify the presence of modified amino acids, the peptide sequence order is still lost and consequently the method cannot indicate the exact site of modification within the measured protein.
A nanopore-based approach for identifying peptide sequences. (A), Peptides are hydrolyzed using an aminopeptidase to produce amino acids, which are passed through a membrane-bound nanopore with a nickel metal ion sensor. The sensor creates an amino acid-specific change in resistance, which can be measured as a local change in current across the membrane. This process loses sequence ordering but retains the amino acid frequencies in the peptide. The accuracy of these frequency measurements changes how reliably the signal can be assigned to a specific peptide in the human proteome. (B), Assuming prior tryptic digestion, the uniqueness of a frequency-based peptide signal from the MspA nanopore (Full Frequency) approaches that of perfect sequence knowledge. Assignment accuracy remains high if at least 2 copies of an amino acid in the peptide produce a distinguishable signal from a single copy but no other higher frequencies can be determined (Limited Frequency). However, knowing just the presence or absence of amino acids in a peptide is insufficient to identify it. (C), With even limited frequency knowledge, almost 90% of proteins can potentially be assigned at least 10 unique peptides between 7 and 30 amino acids in length (the range typically measurable by tandem mass spectrometry).
As part of validating the specificity of the MspA–NTA-Ni mutant, Wang et al. demonstrated that nanopores could rapidly and cost-efficiently assess the composition of commercial amino acid tablets. The authors speculate that since the nanopore was tolerant to nonamino acid components of these tablets (e.g., starch, inorganic salts), it may also be possible to analyze amino acid levels in serum, plasma, or urine. While much work remains to validate if an MspA-based nanopore could be used with these biofluids, the potential clinical applications are enticing. For example, nanopore-based amino acid analysis kits could produce nearly immediate and inexpensive diagnoses for common amino acid metabolic disorders. Phenylketonuria newborn screening is routinely performed within 1 to 3 days after birth using either a fluorometric method or tandem mass spectrometry to measure the amino acid phenylalanine. Replacing these assays with amino acid-agnostic nanopores could simplify early screening for other rare metabolic diseases such as cystinuria.
However, several challenges remain before such nanopores can be applied to general proteomic applications. Unlike more traditional proteomic methods, amino acid-specific nanopores rely on detecting peptides based not on their sequence but on the observed frequency of different amino acids in a peptide. How confidently can we detect peptides using only the frequency of amino acids? Fig. 1B shows that most peptides can be uniquely assigned to a protein from their amino acid frequency alone, and the exact sequence order is generally only necessary to differentiate peptides with fewer than 10 amino acids. However, without a fixed-feed ratcheting mechanism, the measurement of this frequency is potentially challenging. Both the constriction site and the outer rim of MspA are highly negatively charged, which will change the rate at which certain amino acids pass through the pore. Even within a single amino acid type, Wang et al. demonstrate highly variable timing where the event timing (ton) is not considered a predictive feature for each type of amino acid. If amino acid-specific nanopores can only detect the presence (or absence) of amino acids, there are far too many alternative matches (different peptide sequences with the same types of amino acids) within the human genome to make meaningful measurements. However, if nanopores can confidently measure a 2-fold increase in an amino acid, then it may be possible to differentiate most peptides even with just this limited quantitative information. Furthermore, nearly all proteins have sufficient numbers of peptides with unique frequencies to be detectable (Fig. 1C). Future experiments are necessary to test the quantitative capabilities of this type of nanopore.
Nanopore protein sequencing will need to compete meaningfully with other, more established proteomics methods. For over 2 decades, mass spectrometry has been the gold-standard technique for measuring complex mixtures of proteins, peptides, amino acids, and other metabolites. When coupled with liquid or gas chromatography, mass spectrometry is ideally suited to measuring biopolymers with a wide variety of potential modifications due to its ability to separate analytes with high resolution, dynamic range, and quantitative precision. To compete with mass spectrometry, new technologies such as nanopore protein sequencing will need to confidently make hundreds of thousands of quantitative peptide measurements an hour across 5 orders of magnitude. Here, measuring single molecules has a toll: to observe even a single read per peptide at the bottom of that dynamic range, the nanopores will need to be parallelized such that 10 million reads can be made per sample. Furthermore, accurate peptide quantification will require an additional magnitude of reads to differentiate from background signals.
There are reasons to consider moving away from mass spectrometry for proteomics, especially in the clinical space. Mass spectrometry is exceptionally expensive, non-portable, and technically challenging to perform, using hardware that is difficult to maintain. These challenges may be addressed by nanopore sequencing. Single-molecule nanopore protein sequencing synergizes well with low-input proteomics challenges (e.g., single-cell proteomics), since proteins cannot be molecularly amplified. Meanwhile, alternative affinity-based technologies, such as SomaScan (SomaLogic) and Olink, capitalize on coupling high sensitivity with massively parallel DNA-based readouts, making them attractive for measuring proteins in high-background biofluids like plasma and urine. It remains to be seen if these technologies will be able to scale to make reliable global proteomics measurements in clinical settings. Regardless, access to a broader range of proteomics tools will provide researchers and clinicians with more opportunities to make meaningful discoveries.
Questions about the future of nanopore-based proteomic technology now revolve around when it will arrive and what form it will take, rather than if it is possible. With the successful discrimination of all 20 proteinogenic amino acids using a single nanopore, the remaining challenges in front of nanopore protein sequencing are largely technical, engineering-based problems. The small footprint and low cost of this rapidly improving technology have the potential to radically change how proteomics measurements are made and—maybe more importantly—who can make them.
Author Contributions
The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list.
Brian Searle (Conceptualization-Lead, Funding acquisition-Lead, Investigation-Lead, Writing—original draft-Lead)
Authors’ Disclosures or Potential Conflicts of Interest
Upon manuscript submission, all authors completed the author disclosure form.
Research Funding
This research was supported by NIH R35GM150723 to B.C. Searle.
Disclosures
B.C. Searle is a founder and shareholder in Proteome Software, which operates in the field of proteomics. B.C. Searle received grants NIH R01GM133981 from NIGMS, NIH R21CA267394 from NCI, NIH U19AG065156 from NIA, and Thermo Fisher Scientific, received consulting fees from Bruker Daltonics, holds stock in Proteome Software and Talus Biosciences, and participated on the Boards of Bruker Daltonics and Talus Biosciences.
Acknowledgments
We thank D. Wilburn for insightful discussions and J. Haskin for editorial assistance.
