This column highlights recently published articles that are of interest to the readership of this publication.
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.
Simmons et al. document the performance of new, massively parallel, short-read nucleic acid sequencing methodology recently developed by Ultima Genomics, Inc., Newark, CA. The goal of the development effort has been to streamline existing short-read sequencing methods to provide comparable performance in high-throughput applications at substantially reduced cost in consumables. Instead of conventional flow-cells incorporating microfluidic channels, Ultima’s sequencing platform uses circular 200-mm silicon wafers as a substrate. These wafers are standard items in the semiconductor industry. They electrostatically bind clonally amplified sequencing beads at high density. The beads are separately produced by an automated process involving clonal emulsion polymerase chain reaction. Reagents are rapidly delivered to the wafer from nozzles at its center, while the wafer rotates to distribute these reagents rapidly and economically over its surface. Optical signals are recorded during rotation in a continuous process analogous to reading a compact disk. The sequencing chemistry is an implementation of sequencing-by-synthesis. It employs a mixture of fluorescently labelled (terminated) and unlabeled (non-terminated) forms of each nucleotide. The abundance of the non-terminated form of the nucleotide is favored (80%) over the terminated form (20%). The resulting synthesized DNA is therefore mostly unmodified, but there’s enough fluorescent signal to reliably quantify whether 0, 1, 2, or more of each nucleotide has been added to the growing chain in each synthesis cycle after the addition of each of the 4 different nucleotides. This process, termed ‘mostly natural sequencing-by-synthesis,’ combines high throughput with accuracy. Each sequencing cycle takes about 2 min, and the authors can reliably identify SNPs and indels in homopolymer sequences up to 10 residues. Finally, for base calling, the methodology uses graphics processing units to extract bead locations (clonally amplified sequencing units) from the serial images acquired over multiple sequencing cycles. Local, sequence-dependent variation in label incorporation is accommodated using a deep convolutional neural network. The first implementation of this methodology yielded ∼10 billion reads per run, with 300-bp reads, a turnaround time of 20 h/run, and base quality similar to existing platforms (Q30 > 85%). The consumable costs totaled just $1/Gb. Simmons et al. here demonstrate the capabilities of this methodology for large-scale applications in 4 diverse single-cell RNA sequencing case studies. Comparison of the sequence output with Illumina output shows that the two platforms perform similarly. Some detailed differences are noted, as, for example, in positional coverage relative to annotated gene boundaries, presumed to result from Ultima’s use of single-end reads whereas Illumina uses paired-end reads. Significant reduction in sequencing costs being implemented by Ultima, among other vendors, is expected to have a major impact on the scope of DNA sequencing in clinical applications requiring single-cell sequencing, in which very high throughput is necessary.
The T2T-CHM13 reference standard genome recently completed by the Telomere-to-Telomere (T2T) consortium incorporates a contiguous sequence for each human autosome, plus the X chromosome. It constitutes a new gold-standard genome assembly for alignment of other sequences. Yet the extent of human sequence variation precludes high-confidence alignment of many sequences with such a standard, based as it is largely upon sequences from a single individual. Particularly difficult to align are structural variants, which are the focus of increasing functional interest. This problem results in a “reference bias” against the mapping of highly variable genomic regions. The Pangenomic Reference Consortium now announces a new, complementary reference standard that incorporates assemblies from individuals of diverse ancestry. This first draft pangenome includes phased, diploid assemblies from 47 individuals (a total of 94 haploid genomes), based on data that include long-read, as well as conventional short-read sequences at high depth of coverage. To ensure the accuracy of the assemblies – crucial for fulfillment of the project’s goals – the authors deploy recent improvements in de novo assembly and a new pipeline for mapping long-reads to validate the predicted haplotypes. Compilation of alignments for the various individuals into a single pangenome is achieved with a graph-based system of branching and merging paths. The Consortium is working on improvements to reference-based sequence mapping workflows to help replacement of existing references by the pangenome. They note that a substantial number of structural variants can be genotyped using pangenomic methods even on the basis of short-read sequences, although of course, long reads will continue to provide a major advantage. They plan to expand the pangenome to include a cohort of 350 individuals and to work toward providing T2T genomes for each.
Mangiameli et al. describe methodology for sequencing DNA in fixed tissues in spatial regions of interest localized by microscopy, so that they can next be illuminated to activate sequencing in the illuminated area. The authors perform Assay for Transposase-Accessible Chromatin by sequencing (ATAC-seq). ATAC-seq uses a tagmentation process in which sequencing libraries are created in situ using a hyperactive Tn5 transposase, which cleaves DNA in genomic regions of accessible chromatin and adds sequencing adaptors. To accomplish photoselective sequencing, tagmentation adaptors that are blocked to amplification by their conjugation to a fluorophore. Conjugation is via a photocleavable spacer. After microscopic identification of regions of interest in stained tissue, the fluorophores are cleaved from the blocked adaptors by selective illumination with a 404 nm laser. This reveals a 5’ phosphate group to which secondary, indexed adaptors containing a priming site for library amplification can be ligated for subsequent Illumina sequencing. This methodology is validated by profiling chromatin accessibility in oligodendrocytes located in various different regions of mouse brain. The authors then demonstrate use of the procedure at the subcellular scale. They perform selective characterization of chromatin at the nuclear periphery of cultured human fibroblast (IMR90) cells. The present implementation of the methodology is limited by unsuitability for applications requiring 3-D photoselection, because the targeting illumination laser passes through the entire thickness of a specimen and may deprotect fragments above and below the focal plane. This limitation may be surmounted in the future by two-photon absorption, hopefully enabling fully volumetric photoselection.
DelRosso et al. conduct a systematic, high-resolution survey to map the effector domains of the known human transcription factors and chromatin regulators. Effector domains serve to activate or suppress transcription when bound to a DNA target. Whereas DNA binding domains are easy to delineate, effector domains are more difficult because many of them are disordered and poorly conserved. The authors synthesize a library of DNA sequences that encode 80-amino acid segments that tile across 1,292 transcription factors and 755 chromatin regulators with a step size of 10 amino acids. The library is cloned into a lentiviral vector and expressed as fusion proteins with a doxycycline-inducible DNA binding domain, rTetR. The pooled library is expressed in K562 cells at an infection rate that ensures infection of only 1 tile per cell. Expression levels are measured with a 3xFLAG tag on each fusion protein. The cells contain a reporter gene that includes rTetR binding sites for recruitment of effectors. The reporter produces a fluorescent protein for quantification by flow cytometry, and a surface marker for magnetic bead separation of cells. Reporter synthesis is under a minimally active promoter for identifying activators, or a constitutively active promoter for identifying repressors. After magnetic separation into ON and OFF cells, the effectors enriched in each cell population are identified by sequencing. With this methodology, and subsequent validation studies, the authors assign effector domains to 1,568 (77%) of the proteins screened. Rational mutagenesis and deletion scans further help identify residues necessary for effector function. Repressor domains frequently contain motifs that indicate mechanism of action. Mechanisms include SUMO interaction or SUMOylation sites, co-repressor binding motifs, zinc fingers, DNA-binding domains, and dimerization domains. Activation domains manifest significant compositional biases. The results of this study facilitate prediction of transcriptional effector domains, and the methodology may be used in future work in different cell types and signaling conditions to determine how effectors work in varying contexts.
Biological molecules of many different types, including glycans, lipids, amino acids, nucleotides, metabolites and many pharmaceuticals, exist in isomeric forms, with distinct biological characteristics, that share the same exact mass. Occasionally, isomers may be distinguished in product ion spectra by the presence of fragment ions of different mass, but even when such different fragments are unavailable, the relative intensities of product ions may differ because preferred fragmentation pathways may depend upon the structure of the precursor’s transition state. Wu et al. here provide a statistical framework for assessing the confidence with which observed quantitative differences in relative intensities of fragment ions may be used for identification of isomeric species. In the setting of quantitative variability between spectra of the same analyte, their methodology enables calculation of a statistical probability that spectra differing quantitatively in fragment ion intensities actually derive from structurally different analytes. In tests of peptides with various amino acid substitutions, they show a surprisingly high rate of success in distinguishing Leu/Ile, Asp/IsoAsp, and D/L amino acids. The methodology works both with analytes introduction by on-line liquid chromatography and wth continuous infusion. It works with a wide range of instrument settings, and with diverse fragmentation techniques, including collision-induced dissociation, higher-energy collisional dissociation, electron transfer dissociation, and radical-directed dissociation. Because isomeric mixtures are frequently encountered, the authors show that when isomers can be distinguished, the relative abundance of the different isomers can be quantified with suitable calibration curves. The authors also draw attention to possible application of their methodology in the comparison of mass spectra in other contexts, such as normalization of spectra acquired with different instruments, and evaluation of the influence of structure on dissociation.
Bradman et al. demonstrate improvements to the use of multiple stages of tandem mass spectrometry (MSn) for structural characterization of lipid metabolites during LC separation. MS3 scans can be triggered in a data-dependent manner upon detection of a product ion or neutral loss characteristic of a lipid class of interest, but this approach results in the acquisition of a non-trivial number of MSn scans triggered by artifacts. This, of course, diminishes sampling depth without improving structural characterization. Here, the authors utilize a real-time search capability, recently implemented in Orbitrap Tribrid systems, to improve on the information available for data-dependent triggering of class-targeted MSn of lipids. They apply this approach to analysis of phosphotidylcholines, phosphotidylethanolamines, phosphotidylinositols, phosphotidylglycerols, phosphotidylserine, and sphingomyelins. The analyses are performed in the positive ion mode, without recourse to polarity switching, which would increase duty cycle duration or involve additional LC-MS runs. Similar procedures to that of Bradman et al. may be used in varied applications, such as real-time elucidation of sn positional isomers, localization of double bonds, and analysis of other lipid classes.
Ben Falch et al. demonstrate how infrared spectroscopy may be used to complement mass spectrometry in on-line identification of isomeric or isobaric metabolites in the gas phase. Their approach utilizes cryogenic infrared (IR) spectroscopy, in which low temperature increases spectroscopic resolution, and thus facilitates discrimination of subtly different molecules. Gas-phase analyte ions are introduced into a cryogenic ion trap containing a He:N2 mixture (80:20), which is cooled to 45 ˚K by a closed-cycle cryostat. Under these conditions, analyte ions form weakly bound clusters with N2 molecules upon collision. Upon irradiation of the tagged ions with a continuous-wave, mid-IR laser, N2 molecules dissociate when the cluster absorbs a photon whose frequency permits intramolecular vibrational redistribution of energy. The m/z of the trapped ions is then measured by mass spectrometry, in this case in a time-of-flight analyzer. An IR spectral fingerprint is obtained by recording m/z of tagged ions divided by the sum of the signals corresponding to tagged and untagged ions as a function of laser wavenumber in the range 3300-3750 cm-1. This range probes OH and NH oscillators. By appropriately shortening the wavelength scan range and down-sampling the recorded spectra, reproducible fingerprints are obtainable in as little as 7 s. Spectra are compared with a database of analytical standards using a pattern recognition algorithm based on principal component analysis and clustering. The rate of IR spectrum acquisition in this pilot study represents a significant improvement on rates generally associated with IR spectroscopy, and begins to approach rates required for on-line liquid chromatography-tandem mass spectrometry (LC-MS2).
Presently, most drugs are designed to interact with chosen protein targets in pre-defined ways to produce desired cellular responses and therapeutic effects. In this setting, it is perhaps surprising that drugs may produce many unintended cellular effects, or produce therapeutic responses by unintended mechanisms of which we may be unaware. Utilizing drug-regulated post-translational modifications as markers for the engagement of drug targets and pathway recruitment, Zecha et al. illustrate how a proteome-wide screen for changes in post-translational modifications may be used to characterize the actions of drugs. For each drug, they quantify post-translationally modified peptides over a range of drug concentrations at serial time-points. In this way, they acquire detailed dose-response and kinetic profiles of the drug’s action. This information helps identification of the pathways recruited and estimation of their contribution to the physiologic responses elicited. From proteome digests, the authors perform LC-MS2, selecting for analysis acetylated or ubiquitylated peptides by immunoprecipitation, or phosphopeptides by immobilized metal affinity chromatography (IMAC). Multiplex quantification is performed by stable isotope labeling with tandem mass tags (TMT). The authors collect data on 31 anti-cancer drugs belonging to 6 different drug classes, and test these drugs on 13 human cancer cell lines. The methodology enables acquisition of large amounts of information, estimated to represent 1.8 million dose-response curves. Quantification is performed for approximately 47,500 regulated phosphopeptides from 12,000 proteins, 7,300 regulated ubiquitylated peptides from 9,200 proteins, and 550 regulated acetylated peptides from 1,400 proteins. The analysis is facilitated by the finding that most peptides detected to bear these modifications are not regulated by most drugs. The results of the study demonstrate how the methodology can illuminate mechanisms of drug action, and indicate further applications in analysis of resistance mechanisms. The methodology may be further extended to analysis of signals affecting diverse cellular states.
The extracellular matrix (ECM), comprising collagens, fibronectins, laminins and additional glycoproteins and proteoglycans, modulates cell behavior through cell signaling. Although the composition of the ECM varies between tissues, the extent of this variation, and hence its biological effect on the behavior of resident cells, are poorly characterized. Presumably, the insolubility of the ECM components under conditions generally used for proteomic analysis, the large dynamic range of their relative abundance, and their extensive post-translational modification all contribute to the incompleteness of our knowledge. McCabe et al. here perform label-free quantification by LC-MS2 analysis of extracellular matrix preparations from 25 different mouse organs. They cryo-mill and lyophilize tissues, then extract with a CHAPS/NaCl buffer, solubilize in guanidine hydrochloride, and digest with hydroxylamine hydrochloride at pH 9.0 to cleave polypeptides at asparaginyl-glycine linkages. ECM components enriched and solubilized in this way are digested with trypsin for LC-MS analysis. The results confirm that tissues differ in the relative quantities of the major ECM components. Further improvements in methodology to identify and quantify components presently missing from the analysis, such as elastin, and to avoid chemical modifications caused by incubation at high pH, are hoped to provide information with which cell-matrix interactions can be further investigated in organoid and 3-D cell culture.
Genome-wide association studies (GWAS) identify gene loci at which common variant alleles in a population affect characters of interest. The effects identified are usually small, and, even in aggregate, account for only a minor portion of the character’s genetic variance in the population. But they may be useful in providing clues to the character’s physiology or pathology. Discernment of these clues, however, is commonly impeded by the localization of loci identified in GWAS to non-coding regions of unknown function. The present paper provides a systematic approach to discovery of the function of such loci. The authors investigate many quantitative characters of blood cells commonly measured in clinical practice (e.g. mean corpuscular volume, leukocyte count), each of which is affected by many gene loci (i.e. they are polygenic). The authors search for candidate cis-regulatory elements (cCREs) associated with GWAS loci, and their putative target genes. They perform fine-mapping of GWAS loci, then target putative cCREs in pooled CRISPRi silencing screens using a strong repressor system in the hematopoietic cell line K562. Cell responses are tested by monitoring CRISPR guide RNAs, transcriptome expression (including non-coding RNAs), and cell surface expression using oligonucleotide-tagged antibodies with ‘cell hashing’. Fine mapping by silencing of multiple putative cCREs at the same GWAS locus enables causal variants to be distinguished from non-causal variants in linkage disequilibrium with them. A subset of cCREs identified in this way is further investigated in single-cell, pooled base editing screens. The authors also identify trans-CREs for GWAS loci, including some for which cCREs are also identified. The authors successfully identify target genes for 25% of the cCREs they test, and 36% of the GWAS loci they study. As high as these yields are in comparison to earlier protocols, the authors caution that the results do not exclude the likely existence of other causal variants, CREs and target genes, and recognize that effects of GWAS loci in primary cell types may remain to be discovered.
When CRISPR nucleases introduce a double-stranded break in cellular DNA at a target site specified by a guide RNA, the cell initiates DNA repair. The experimenter may direct insertions, deletions or substitutions to occur during repair at the site of the break by supplying an exogenous DNA donor of the desired sequence. Unfortunately, the cell frequently introduces other sequence changes at the target site that are undesired by the experimenter, e.g. larger deletions, or gene conversion to the sequence templated by the homologous chromosome. Either of these particular undesired changes results in loss of heterozygosity. Detection of such undesired, on-target effects can be difficult, but Lackner et al. suggest a streamlined strategy for their identification upon sequencing the target site in clones that have undergone gene editing. They supply as the donor oligonucleotide a mixture of oligonucleotides of length 90 nt, whose components, in addition to the desired operational replacement at the target site, also specify various substitutions that make no difference to the function of the edited allele, but indicate that editing has occurred when detected by later sequencing. Because a mixture of donor oligonucleotides specifying different “diagnostic” substitutions of this kind are present, when both chromosomes in a homologous pair undergo editing, the two chromosomes will usually incorporate different diagnostic substitutions from different oligonucleotides in the mixture, so the fate of the two partners can be followed in the heterozygous state. The expected outcome upon sequencing is the presence of the desired replacement in homozygous state (indicating that both homologs have undergone editing), and the diagnostic substitutions in heterozygous state (indicting that unintended effects have not occurred). The absence of heterozygous diagnostic substitutions would indicate that unintended changes have occurred on one of the homologs. An abnormal number or ratio of the diagnostic substitutions would suggest possible multiple insertion of the donor sequence, or duplications, or aneuploidy of the target chromosome, or a polyclonal cell population with various different editing effects. The authors validate this scheme in a study involving editing of 850 human embryonic stem cell clones with such donor mixtures.
This study addresses the prevalence of eukaryotic DNA viruses to be expected in metagenomic studies of human tissues from individuals not experiencing viral illness. The authors test tissue samples for DNA from 38 viral species using both quantitative polymerase chain reaction (qPCR) and, for viral enrichment, hybrid capture high-throughput sequencing. The test samples represent colon, liver, lung, heart, brain, kidney, skin, blood and hair from 31 recently deceased, Finnish individuals aged 36-85 years. The authors identify a total of 17 viral species in these samples, with an average of as many as 6.7 species per individual. Lung, liver, kidney and colon show the largest number of species (3.6-3.9 per organ), and brain the smallest number (average 1.2). These are surprisingly high numbers compared to expectations based on body fluids, which are the more commonly studied. Viral genome coverage in the study is generally high, with a total of 70 genomes assembled at > 90% breadth of coverage. The distribution of viruses among tissues varies both qualitatively and quantitatively among viral species, but is generally low, with an estimated mean of 540 copies per million cells. Such low levels raise concern about possible contamination from non-tissue sources of viral DNA during processing. The authors employ negative controls to test for this throughout, and observe that varying strain identities and prevalence, as well as distinctive viral content of different tissues, all render contamination an improbable determinant of the results. The study’s findings complicate attempts to make causal connections between diseases and sequence-based identification of viruses. The results also raise questions about the safety of tissue transplantation, especially under conditions that risk viral reactivation.
In April 2022, there began an outbreak of acute, severe, pediatric hepatitis of unknown cause that has now affected over 1,000 children in at least 35 countries. The work of 3 independent groups to identify the cause of the disease illustrates the complexity involved in identifying pathogenic agents. Previous studies based on targeted sequencing had associated the outbreak with human adenovirus (HAdV), but the present 3 groups show a link with adeno-associated virus 2 (AAV2). This virus had not previously been known to cause hepatitis, although AAV is commonly used in gene therapy trials, in which hepatitis has been observed, although infrequently. Ho et al. use untargeted high-throughput sequencing, and PCR with reverse transcription, to show the presence of AAV2 in 26 of 32 cases (81%) with hepatitis, compared to 5 of 74 (7%) of samples from unaffected individuals. They also note, by in situ hybridization, the presence of AAV2 RNA in cellular settings indicative of active viral replication. Data from patients’ sera do not indicate a robust humoral immune response specific for AAV2, but, in the liver, elevated numbers of CD68+ macrophages, activated CD4+ and CD8+ T cells, and CD20+ B cells indicate strong localized immunologic activity. The authors comment that 25 of the 27 affected children (93%) bear the HLA class II allele HLA-DRB1*04:01 compared with 10 of 64 controls (16%), indicating genetic susceptibility to the disease. Morfopoulou et al. use similar methods and report similar findings. However, they also note results of PCR tests showing, in addition to high levels of AAV2 DNA, low levels of human adenovirus (HAdV) in 23 of 31 cases (74%), and of human herpesvirus 6 (HHV-6) in 16 of 23 cases (70%). They provide evidence for active replication and transcription of AAV2 DNA, and suggest the possibility that HAdV and/or HHV-6 may act as helper viruses upon which replication of AAV2 in the hepatitis cases depends. These authors also note that the sequences of AAV2, HAdV and HHV-6 exclude the emergence of novel strains in the context of the cases. Importantly, both immunohistochemistry and proteomic methods fail to detect AAV2 proteins, indicating that viral particles are not being formed in these patients. Servellita et al. restrict their analyses to hepatitis cases that are positive for HAdV. Among 14 such cases, the blood of 13 (93%) is positive for AAV2, compared to 3 of 113 (3.5%) of controls. Co-infection with Epstein Barr virus is detected in 11 of these 14 cases (79%), and HHV-6 in 7 (50%), again suggesting the possibility that such viruses may be acting as helper viruses in AAV2 replication. Finally, all 3 groups note how the hepatitis outbreak has occurred in the setting of the COVID-19 pandemic. SARS-CoV-2 virus may have contributed to hepatitis pathogenesis directly, but, because the hepatitis outbreak peaked as lockdown measures relaxed, it may have affected exposure to many viruses in immunologically vulnerable children. These studies illustrate the multiplicity of techniques required for elucidation of the causes of viral illness.
The gram-negative bacterium Photorhabdus asymbiotica, a pathogen of insects that causes them to glow in the dark once infected, has a genome that encodes a contractile injection system, a syringe-like molecular complex that delivers toxins into eukaryotic cells. Tail-fibers in the complex recognize a receptor on a target cell membrane. Once the complex is anchored to the target cell surface, contraction of an outer sheath forces an enclosed, rigid, sharpened tube through the cell membrane. A payload of proteins contained within the complex is then delivered into the cell upon disassembly of the spike-tube structure. Kreitz et al. here show that this delivery system can be re-engineered to target specific cells, and to deliver customizable payloads – a programmable protein delivery device. By analogy with the structure of bacteriophage T4, they deduce which gene encodes the protein conferring target cell specificity. They use AlphaFold to predict the 3-D structure of the likely binding domain, then insert novel sequences with known specificity for mouse or human cell surface proteins to alter the tropism of the complex. They further deduce that a disordered region of the native payload protein represents a packaging domain, and fuse it to novel proteins to construct novel payloads. The resulting constructs deliver novel cargos to newly targeted cells with high efficiency. The authors demonstrate selective killing of cancer cells, and the delivery of the large protein Cas9 that mediates gene editing when supplied to cells harboring a guide RNA. The injection system is active in vivo in mice with little activation of immune cells, inflammation, weight loss or cellular toxicity, indicating low immunogenicity and toxicity. These results suggest diverse possible therapeutic and physiologic applications for this novel injection system.
The capability to observe motions of individual macromolecules within the cellular environment on a millisecond timescale, unencumbered by conjugation to large beads, has long been an aspiration of optical microscopy. This goal is achieved in the present pair of papers, and the methodology is applied to observation of the stepping movements of kinesin. The authors build upon the recently described technique named MINFLUX. In this technique, a fluorophore is localized not by the obvious means of detecting where its emission is brightest, but rather by illuminating it with a donut-shaped illumination spot within the area of a few hundred nanometers where the fluorophore is located, then moving the donut around to locate where the emission intensity of the fluorophore is at its minimum, the location that coincides with the absence of excitation in the donut’s central hole. The advantage of this approach is that it requires 10 to 100 times fewer photons to locate the fluorophore than comparable resolution by conventional camera-based techniques, thereby minimizing photobleaching. Previous work with MINFLUX, like the present study, employed photo-switchable dyes in which the fluorophore can be switched between a fluorescent state and a dark state by illumination with different wavelengths. They excite only a small subset of fluorophores at a time, allowing their localization individually. This protocol demonstrated that this methodology can achieve spatial resolution down to 2-3 nm; i.e. true molecular-scale imaging. Wolff et al. now describe a new implementation of MINFLUX in which interferometry is used instead of a donut pattern to create an even more sharply localized illumination intensity minimum that further improves the precision with which a fluorophore can be localized. They achieve the recording of protein movements with spatial precision of 1.7 nm (17 Å) and temporal precision of 1 ms for prolonged periods. This enables tracing of motions of the 2 kinesin heads as fast as they occur at physiologic concentrations of ATP (i.e. not slowing them down by reducing ATP concentration). This allows details of intramolecular conformational motions to be recorded during the molecule’s duty cycle. Deguchi et al. extend the analysis by tracking the motion of kinesin in live cells at comparable spatiotemporal resolution as it walks along microtubules. The authors look forward to developing two-color tracking to monitor the relative motions of different molecules or molecular domains.
In this paper, Weber et al. use a different optical method for achieving super-resolution (i.e. resolution below the diffraction limit). The method is based on stimulated emission depletion microscopy (STED), in which a fluorophore is localized by a Gaussian excitation beam, but the emission signal is confined by a second, coaxial, donut-shaped excitation beam with wavelength tuned to simultaneously de-excite the fluorophore by stimulated emission. In the technique of MINSTED, only a small subset of fluorophores is switched on at any one time, so the reduced background minimizes the required number of fluorescence detections required for STED microscopy, and therefore minimizes photobleaching. Resolution enhancement occurs because the center region, from which the camera gathers fluorescence radiation, is much smaller than the diffraction-limited area. This technique has achieved a precision of 1-2 nm in the localization of fluorophores, but further improvement has been limited by heating of the sample and lens immersion oil by the STED beam. Here, the authors find they can blue-shift the donut STED beam without incurring direct excitation (and higher background) from nearby fluorophores as a result of straying into their emission wavelengths, because so few fluorophores are switched on in MINSTED. Blue-shifting permits a much lower STED beam power to be used, and heating is thereby avoided. The result is that fluorophores can be localized to within 0.47 nm (4.7 Å) – much smaller even than the diameter of the fluorophore itself. The authors apply this capability to imaging of nuclear core complexes and nuclear lamins. This technology opens up possibilities for imaging macromolecular motions in the Ångstrom spatial range for exploration of molecular dynamics within cells.
Brillouin light scattering is an inelastic process that results from the interaction between light, usually from a laser, and spontaneous, thermally induced density fluctuations in a material. The density fluctuations can be described as a population of microscopic acoustic waves, or quasi-particles called phonons, which represent the collective, periodic excitation of molecular densities in the material. Light scattered elastically has the same frequency as the incident light (Rayleigh scattering), but a small fraction (∼10-12) can interact with phonons and exchange energy and momentum with them, resulting in Doppler shifting to produce positive or negative changes in the frequency of the inelastically scattered light. Analysis of the Brillouin spectrum therefore provides information about the material’s mechanical properties because sound wave velocity or attenuation are affected by the material’s elasticity and viscosity. The magnitude of the light’s frequency shift provides information about the material’s elasticity, and the broadening of its frequency provides information about viscosity. The viscoelastic characteristics of cells are strongly affected by their internal structure and state of hydration, and the viscoelastic characteristics of tissues is affected additionally by the structure and hydration of the extracellular matrix. Because Brillouin signals are weak, acquisition of the spectra has required high intensity of illumination and long acquisition times. But two groups now describe improvements to microscopic methods for acquisition of Brillouin spectra that alleviate these limitations, enabling expanded application of Brillouin microscopy in cell and developmental biology. Both groups adopt line-scanning to enable multiplexed signal acquisition, allowing simultaneous sensing of hundreds of points in a sample and acquisition of their spectra in parallel with low phototoxicity. Zhang et al. apply their technique to the study of tumor spheroids, and Bevilacqua et al. study developing Drosophila, ascidian and mouse embryos. It is hoped that continuing developments in this field will provide expanding capabilities for investigation of biomechanical properties of cells and tissues relevant to developmental biology and pathology, with applications in tissue engineering and clinical diagnosis.
Microbial communities are presumed to vary in composition and activity along the length of the gastrointestinal (GI) tract in ways that both adapt to, and modify, local environmental conditions. These variations cannot be adequately investigated either by stool sampling, or by invasive procedures such as endoscopy, during which the microbiome may be significantly perturbed. In order to investigate normal variation in the microbiome along the length of GI tract, and its contribution to digestion and absorption, Shalon et al. develop a capsule device consisting of a collection bladder capped with a one-way valve for retrieval of a 400-µL liquid sample. Bladders are packaged inside a dissolvable capsule. Different enteric coatings of the capsule are designed to disintegrate at a defined target pH values for sampling from regions proximal or distal to the stomach, with one also incorporating a time-delay to sample from the ascending colon. The different coatings distinguish clearly the region of the duodenum and jejunum from the region of the ileum and ascending colon. Using metagenomic analysis, the authors document microbial communities that are spatially distinct in taxonomic composition and functional capability, and are also different from stool. Bile acid concentration profiles quantified by LC-MS display gradients of microbially induced bile acid modifications that are associated with the abundance of particular bacterial species. Their functional significance remains to be determined. Folz et al. utilize the sampling methodology for analysis of metabolites, including the products of digestion and absorption of dietary components and products of microbial metabolism. Anticipated applications of the methodology include study of the effects of antibiotics and of variation in dietary and host factors, both physiologic and pathologic.
The use of RNAs as therapeutic agents has leaped to prominence during the response to the COVID-19 pandemic, and development of methods for their design and deployment has secured high priority. In this setting, interest has sharpened in circular RNAs (circRNAs), single-stranded RNA molecules in which the 5′ and 3′ ends are covalently joined, in the hope of turning their resistance to degradation by exonucleases to advantage in augmenting the yield of proteins circRNAs are capable of encoding. The mechanisms by which circRNAs are translated differ from mRNAs, notably in the way they recruit initiation factors, and in their reliance on internal ribosome entry sites (IRES). But the details have hitherto been sparsely investigated. The present study identifies and optimizes key facets of circRNA structure affecting translation efficiency, and improves circRNA protein yields by several hundred-fold. For the purpose of this study, the authors develop a modular cloning platform for high-throughput assembly of synthetic circRNAs to incorporate numerous sequence variations for independent optimization of many structural features. The platform incorporates several design features for synthesis and product purification that will be of general utility for investigators in the field of RNA synthesis. For optimization of translation efficiency, the authors insert locked nucleic acids (LNAs) to disrupt secondary structure at optimal positions. They incorporate aptamers at optimal positions relative to IRES to assist in recruitment of initiation factors. And they employ sequence shuffling for design of IRES to tune translational activity. The authors demonstrate sustainably high-level production of erythropoietin in mice from circRNA delivered with charge-altering releasable transporters (CARTs). Many new opportunities for further improvements are opened by this study in the translation of transgenes for both basic and clinical investigations.