Skip to main content
SearchLoginLogin or Signup

Article Watch: September, 2022

This column highlights recently published articles that are of interest to the readership of this publication.

Published onSep 29, 2022
Article Watch: September, 2022


This column highlights recently published articles that are of interest to the readership of this publication. We encourage ABRF members to forward information on articles they feel are important and useful to Clive Slaughter, AU-UGA Medical Partnership, 1425 Prince Avenue, Athens GA 30606. Tel; (706) 713-2216: Fax; (706) 713-2221: Email; [email protected] or to any member of the editorial board. Article summaries reflect the reviewer’s opinions and not necessarily those of the Association.


Morgenstern D, Wolf-Levy H, Tickotsky-Moskovitz N, Cooper I, Buchman A S, Bennett D A, Beeri M S, Levin Y. Optimized glycopeptide enrichment method–it is all about the sauce. Analytical Chemistry 94;2022:10308-10313.

The two most commonly used methods for enrichment of glycosylated peptides present in protein digests are hydrophilic interaction liquid chromatography (HILIC) and lectin affinity chromatography. Both methods exhibit undesirable selectivity: HILIC preferentially enriches peptides rich in hydrophilic amino acids, and lectins target specific glycan subclasses, a problem only partially circumvented by the use of multiple lectins. Boronic acids, which covalently bind 1,2- and 1,3-cis-diols, provide an alternative means of unbiased glycopeptide enrichment, but often require inconvenient special measures to achieve high glycopeptide coverage. Morgenstern et al. here improve the efficiency of boronic acid enrichment by optimizing conditions for glycopeptide binding. They perform comparisons of glycopeptide coverage and quantitative yield using alternative binding buffers: triethylammonium acetate or glycine or carbonate at pH 10.5, and ammonium carbonate at pH 8.5. The best performance is observed with carbonate at pH 10.5, the one amine-free buffer, which provides over 10-fold improvement in the total number of quantified peptides. The authors attribute the poorer performance of amine-containing buffers to competing interaction between amines and the boronic acid stationary phase. The authors also incorporate into their optimized methodology buffers having high ionic strength to suppress electrostatic interactions, and 50% acetonitrile to overcome hydrophobic interactions during purification.


Wang J, Lisanza S, Juergens D, Tischer D, Watson J L, Castro K M, Ragotte R, Saragovi A, Milles L F, Baek M, Anishchenko I, Yang W, Hicks D R, Expòsit M, Schlichthaerle T, Chun J-H, Dauparas J, Bennett N, Wicky B I M, Muenks A, Dimaio F, Correia B, Ovchinnikov S, Baker D. Scaffolding protein functional sites using deep learning. Science 377;2022:387-394.

The authors extend the recent success of deep learning approaches to protein design. They tackle de novo design of amino acid sequences that preferentially fold in three-dimensional (3D) space to provide a scaffold which orients the key functional residues determining the specificity of a binding site. The task is to accomplish this design without specifying in advance the folding pattern or the secondary structure of the scaffold. The authors previously showed that the structure prediction neural network trRosetta can perform Monte Carlo sampling in sequence space to design sequences that will fold into definable structures – a process named “hallucination” because the output corresponds to no known protein yet shares with proteins the ability to adopt some definite, predictable structure. The same network can be used to design sequences that fold into a pre-specified target structure. Here the authors demonstrate that trRosetta can perform scaffolding of a functional site, an exercise characterized as “constrained hallucination.” But a newer network, RoseTTAFold, performs even better than trRosetta because of its superior capability for modeling sequence-structure relationships. As a second solution to the functional site scaffolding problem, the authors formulate the task as an information recovery problem, analogous to completing a sentence given its first few words using language models, or completing a corrupted image using inpainting. Beginning with a functional site, a specifically trained RoseTTAFold network fills in sequence and structure to create a protein scaffold that supports the desired site. The network modified for inpainting, called RoseTTAFoldjoint, is less computationally intensive than that required for constrained hallucination: it requires only 1-10 s per design on an NVIDIA RTX 2080 graphics processing unit, whereas hallucination requires 5-20 min. But the hallucination approach gives better results and generates greater structural diversity when the sequence to be designed is large. Both approaches require no input besides the identity and orientation of the particular residues required for functional specificity of the binding site. The authors use the two approaches to design immunogens, receptor traps, metalloproteins, enzymes, and protein-binding proteins, and validate their designs computationally and experimentally.


Wan N, Wang N, Yu S, Zhang H, Tang S, Wang D, Lu W, Li H, Delafield D G, Kong Y, Wang X, Shao C, Lv L, Wang G, Tan R, Wang N, Hao H, Ye H. Cyclic immonium ion of lactyllysine reveals widespread lactylation in the human proteome. Nature Methods 19;2022:854-864.

Core histones have recently been found to undergo posttranslational addition of lactic acid to lysine residues. Lactylation is an epigenetic mark that modulates gene transcription. Its occurrence affects, and is affected by, cellular physiology. Wan et al. now investigate the prevalence of this post-translational modification on proteins other than histones in human cells. Lactylation confers a precursor ion mass shift of 72.02 Da, but the search for the modification is complicated because lactylation of lysine confers resistance to cleavage by trypsin at the modified residue. The authors show that the most reliable indication of lactylation is the presence in product ion spectra of a cyclic immonium ion (m/z 156.10) derived from lactyllysine (m/z 173.13) following collision-induced dissociation. They use the presence of this product ion to identify lactylation sites in existing proteome databases, the Meltome Atlas and the draft map of the human proteome. Their data mining reveals that many glycolytic enzymes are lactylated. Furthermore, enzyme assay of aldolase A shows that the catalytic activity of the enzyme is diminished by the modification, suggesting a possible (although yet to be proven) role for the modification in negative feedback regulation of the glycolytic pathway. The location of lactylation at residues known to bear other modifications, such as acetylation and ubiquitylation, also raises the interesting question of crosstalk between these alternative modifications. This study opens opportunities for quantitative investigation of the occurrence of a new post-translational modification and its physiological impact.


Bergman D T, Jones T R, Liu V, Ray J, Jagoda E, Siraj L, Kang H Y, Nasser J, Kane M, Rios A, Nguyen T H, Grossman S R, Fulco C P, Lander E S, Engreitz J M. Compatibility rules of human enhancer and promoter sequences. Nature 607;2022:176-184.

Martinez-Ara M, Comoglio F, Van Arensbergen J, Van Steensel B. Systematic analysis of intrinsic enhancer-promoter compatibility in the mouse genome. Molecular Cell 82;2022:2519.

Two groups undertake experimental studies of how cis-regulatory elements (CREs) - either enhancers or repressor elements - choose which gene promoters to interact with to effect transcriptional regulation. 3D folding and chromatin compartmentation contribute to the specificity, but so too does biochemical compatibility between CREs and promoters. The present work seeks to quantify the role played by such biochemical compatibility. Both groups employ massively parallel reporter assays in which pairwise combinations of CREs and promoters are cloned into barcoded, plasmid-based reporter vectors. Pooled vector libraries are transiently transfected into mammalian cells, and the rates at which CRE/reporter combinations drive the transcription of the reporter gene are measured by high-throughput sequencing. The use of a plasmid eliminates contributions due to chromatin architecture. Berman et al. test the combinatorial compatibility of 1000 CREs with 1000 promoters chosen from across the human genome based on diversity in expression or activity levels. Martinez-Ara et al. test 556 CREs and 84 promoters from 3 selected regions surrounding the pluripotency genes Nanog, Tfcp2l1 and Klf2, representing a total of ∼60 genes. Both groups quantify expression of the reporter gene relative to expression measured with the relevant promoter alone. Details of experimental protocol, such as the length of DNA constituting CRE and promoter sequences, and placement of CRE relative to the reporter gene, differ between the two groups. The two groups formulate somewhat different interpretations of the data they acquire. Bergman et al. suggest that most enhancer CREs activate all promoters and that the degree of activation is determined multiplicatively mainly by the intrinsic activities of enhancers and promoters. However, they indicate that the promoters of housekeeping genes show somewhat lower responsiveness to variation in activity of different enhancers than do promoters of variably expressed genes. Martinez-Ara et al. emphasize variability between enhancer CREs in their compatibility with different promoters, and indicate that housekeeping promoters tend to have similar CRE preferences while promoters of variably expressed genes exhibit wider diversity in compatibility with different CREs. The two groups have shared and cross-analyzed one another’s data, and find support for their own interpretations in both datasets, suggesting broad underlying agreement. Interpretation of such data in terms of intrinsic compatibility within CRE/promoter combinations depends upon assumed but untested absence of limitations in transcriptional output resulting from availability of transcription factors, adaptor proteins and transcription initiation components in the host cells employed and the levels at which they are transfected. It also depends upon absence of differential effects of inhibitory signals. These studies nevertheless provide a springboard for further investigation of the physical and functional interactions upon which variation in biochemical compatibility between CREs and promoters is based.

Pinglay S, Bulajić M, Rahe D P, Huang E, Brosh R, Mamrak N E, King B R, German S, Cadley J A, Rieber L, Easo N, Lionnet T, Mahony S, Maurano M T, Holt L J, Mazzoni E O, Boeke J D. Synthetic regulatory reconstitution reveals principles of mammalian Hox cluster regulation. Science 377;2022:eabk2820.

Many biological processes are mediated by complex clusters of genes occupying extended genomic regions. Disentangling the way such gene clusters are controlled in the setting of contributions from features of the topologic chromatin organization and distal or intrinsic enhancers and can be daunting, and requires a capability for simultaneous manipulation of multiple regulatory elements across a broad span of DNA. Until recently, engineering of the requisite sequences by gene editing techniques has not been possible. Pinglay et al. instead adopt an approach based upon synthetic biology. They seek to understand the control of the HoxA gene cluster, which encodes transcription factors that contribute to developmental patterning in response to morphogens such as retinoid acid. Specifically, they wish to distinguish the role of regulatory sequences within the cluster and regulatory sequences in distal enhancers for controlling the cluster’s activity. They synthesize variants of the rat HoxA cluster spanning 130-170 kilobases and deliver them to an ectopic site in the mouse genome (the Hprt locus on the X chromosome) regarded as a “safe harbor” for the purpose. The constructs are derived from overlapping polymerase chain reaction (PCR) amplicons recombined in yeast to form the final assembly for insertion. Variants are made by switching wild-type amplicons with synthetic DNA of the desired sequence or by directly editing with CRISPR/Cas9. Using these constructs, Pinglay et al. demonstrate the essential role of retinoic acid response elements within the cluster, but show that distal enhancers are needed to achieve full transcription output of the genes. The synthetic regulatory reconstitution approach is expected to help dissect the regulation of other such complex systems.


Hirano M, Ando R, Shimozono S, Sugiyama M, Takeda N, Kurokawa H, Deguchi R, Endo K, Haga K, Takai-Todaka R, Inaura S, Matsumura Y, Hama H, Okada Y, Fujiwara T, Morimoto T, Katayama K, Miyawaki A. A highly photostable and bright green fluorescent protein. Nature Biotechnology 40;2022:1132-1142.

Both brightness and photostability are desirable qualities in fluorescent proteins used in imaging applications, but there is generally a trade-off between these two qualities. Hirano et al., however, report the cloning of a green fluorescent protein from the jellyfish Cytaeis uchidae into which they introduce a single point mutation that increases brightness without compromising photostability. The resulting protein, called StayGold, matches the brightness of mNeonGreen, yet is >10x more photostable than any fluorescent protein currently available. The authors use the protein in several applications. They express a StayGold gene in mammalian cells for use as a soluble marker of the endoplasmic reticulum (ER) lumen. This requires some genetic engineering to replace certain disulfide bonded cysteine residues with isoleucines, and to add amino acid sequences at the N- and C-termini to target the protein to the ER. But these changes are made without compromising brightness or photostability. The authors use this construct with grazing-incidence structured illumination microscopy (GI-SIM) over extended periods of illumination to document ER dynamics without photobleaching. They also employ StayGold to study mitochondrial fusion and fission during cell stress. This requires engineering with a mitochondrial targeting sequence. For this application they successfully employ GI-SIM at high resolution. Using StayGold for labeling proteins is complicated by the dimeric nature of StayGold, although the authors are developing a monomeric version. Meanwhile, they engineer a tandem dimer in which the two subunits are coupled with a flexible linker sequence, and fuse this construct to microtubule-associated end-binding protein 3 for extended observation of microtubule dynamics. Success in these applications indicates that the new fluorescent protein will provide substantially expanded capabilities for high resolution live cell imaging and volumetric imaging of tissues.


Murray L P, Mace C R. Paper-based cytometer for the detection and enumeration of white blood cells according to their immunophenotype. Analytical Chemistry 94;2022:10443-10450.

Counting of white blood cells (WBCs) and their subtypes has become a mainstay of clinical diagnosis. But the procedure has been standardized in terms of flow cytometry, a technique that involves bulky, expensive instrumentation and requires special expertise for its operation and maintenance. Encouraged by the commercial success of lateral flow tests and paper-based microfluidic devices for point-of-care use (e.g. pregnancy testing and at-home COVID testing), Murray & Mace explore the feasibility of counting WBCs using paper-based assays suitable for deployment in resource-limited settings. The difficulty with transporting WBCs through paper by wicking is that most traditional chromatography papers have pores that are too small to allow unimpeded passage of the cells. Instead, the authors find that coffee filters work very well: Clever coffee paper filters from Amazon allow unimpeded transport of WBCs in both the vertical and lateral directions. The authors pretreat the paper with bovine serum albumin, then spot horse radish peroxidase-conjugated antibody specific for a leukocyte surface antigen. A cell suspension is then spotted in the same location, washed with phosphate-buffered saline to remove unbound antibody, and transported by wicking along a channel in the paper delimited by wax barriers. At their destination, the cells are captured by a polyethersulfone membrane with pore size small enough to cause the cells to stick. The capture membrane is removed, affixed to a plastic tray, incubated with the chromogenic horseradish peroxidase substrate 3,3′,5,5′-tetramethylbenzidine, and the color is quantified in an EPSON Perfection V600 Photo scanner to provide a measure of the number of the antibody-labeled cells present. In a series of proof-of-concept experiments, the authors demonstrate enumeration of select leukocyte populations using representative cultured cell lines, and show that the system can detect abnormal cell counts in clinically relevant ranges. The methodology must now be validated with samples of whole blood. Successful outcome is hoped to pave the way for development of a new class of low-cost, point-of-care cytometers.


Boehnke N, Straehla J P, Safford H C, Kocak M, Rees M G, Ronan M, Rosenberg D, Adelmann C H, Chivukula R R, Nabar N, Berger A G, Lamson N G, Cheah J H, Li H, Roth J A, Koehler A N, Hammond P T. Massively parallel pooled screening reveals genomic determinants of nanoparticle delivery. Science 377;2022:eabm5551.

The use of nanoparticles for delivery of RNA vaccines for COVID has highlighted the potential of nanoparticle-based therapeutics in medicine. Working in the field of personalized cancer therapy, Boehnke et al. are exploring what features of nanoparticle design affect nanoparticle capability for delivery of chemotherapeutic agents to cancer cells. Here they focus on the properties that determine cell targeting and uptake. They assemble a library of 35 different kinds of nanoparticle that vary in core composition, surface chemistry, and size. Because uptake is likely to vary with different characteristics of the cancer cells themselves, the nanoparticle library is tested against a panel of 448 cancer cell lines representing 22 different cell lineages. Fluorescent nanoparticles are employed without cytotoxic cargo, and their uptake is measured by fluorescence-activated cell sorting. Pooled screening is conducted with DNA-barcoded cell lines. The representation of each cell line in cell mixtures categorized according to fluorescence intensity following nanoparticle uptake is measured by high-throughput sequencing. The cellular features that correlate with uptake are discerned using univariate and random forest algorithm analysis. Results are correlated with characteristics of the cell lines recorded in the Cancer Cell Line Encyclopedia (CCLE). The characteristics include cell lineage; gene copy number; mRNA & microRNA abundance; protein and metabolite abundance; and function-damaging, hotspot and missense mutations. Some of the resulting observations are surprising. Although nanoparticle surface chemistry is expected to determine uptake, the authors instead show that core composition is a primary determinant. And, in addition to the expected proteins mediating drug import, e.g. cell surface solute carrier transporter (SLC) proteins and ATP-binding cassette (ABC) proteins, SLC46A3, a lysosomal transport protein, acts as an inhibitor of nanoparticle uptake. These results indicate that the methodology of Boehnke et al. will hasten the design and personalized administration of nanoparticle therapeutics for many applications. The scope of the work is anticipated to be broadened even further by including more kinds of nanoparticle, more correlative characteristics of target cells, and alternative methods for targeted nanoparticle delivery.

Lin K, Gueble S E, Sundaram R K, Huseman E D, Bindra R S, Herzon S B. Mechanism-based design of agents that selectively target drug-resistant glioma. Science 377;2022:502-511.

Malignant cells typically exhibit genome instability. This instability usually arises from defects in one or more of the pathways that normally contribute to the DNA damage response. Paradoxically, such defects can be exploited for selective therapeutic killing of tumor cells by provoking DNA damage that the defective cells are incapable of repairing. This principle is used by Lin et al. to design a new approach to therapy for glioblastoma, a common brain malignancy with dreadful prognosis. In ∼50% of glioblastomas, and >70% of grade II and III glioblastomas, expression of the DNA repair enzyme O6-methylguanine methyltransferase (MGMT) is repressed by promoter methylation of the MGMT gene. MGMT repairs O6-methylguanine DNA adducts by a suicide mechanism, in which the methyl group on the damaged DNA is transferred to an acceptor site on the enzyme protein, with inactivation of the enzyme. A clinical intervention with some benefit for treatment of glioblastoma is administration of temozolomide (TMZ), a prodrug that converts to a strong alkylating agent which causes various forms of methylation. Most of this methylation is readily resolved by base excision repair, but ∼5% of the product is O6-methylguanine, which remains unresolved in MGMT-deficient tumors. In MGMT-deficient tumors, the adducts accumulate, resulting in cell death via the mismatch repair pathway. Unfortunately, most glioblastoma cells eventually become resistant to TMZ by mechanisms that involve loss of the mismatch repair pathway. Lin et al. synthesize an imidazotetrazine called KL-50, which causes the formation of toxic DNA inter-strand cross-links. These cross-links form rather slowly. MGMT forestalls their formation, so the toxicity of KL-50 for cells expressing MGMT is low. But in MGMT-deficient cells the cross-links accumulate and cause cell death by a mismatch repair-independent pathway. Glioblastoma cells that have become resistant to MTZ are therefore killed in this way. The authors note that 40% of colorectal cancers and 25% of non-small cell lung cancer, lymphoma, and head and neck cancers also exhibit MGMT silencing, and several other types of cancer exhibit reduced expression of MGMT, suggesting that KL-50 might also be useful in settings other than glioblastoma treatment. Clinical trials are awaited to test this therapeutic approach. The introduction of DNA lesions that are rapidly repaired in normal cells but not in DNA repair-deficient tumor cells is an approach with potentially widespread applicability.

No comments here
Why not start the discussion?