Nature Biotechnology

Published by Springer Nature
Online ISSN: 1546-1696
Learn more about this page
Recent publications
Article
CRISPR–Cas13 systems have recently been used for targeted RNA degradation in various organisms. However, collateral degradation of bystander RNAs has limited their in vivo applications. Here, we design a dual-fluorescence reporter system for detecting collateral effects and screening Cas13 variants in mammalian cells. Among over 200 engineered variants, several Cas13 variants including Cas13d and Cas13X exhibit efficient on-target activity but markedly reduced collateral activity. Furthermore, transcriptome-wide off-targets and cell growth arrest induced by Cas13 are absent for these variants. High-fidelity Cas13 variants show similar RNA knockdown activity to wild-type Cas13 but no detectable collateral damage in transgenic mice or adeno-associated-virus-mediated somatic cell targeting. Thus, high-fidelity Cas13 variants with minimal collateral effects are now available for targeted degradation of RNAs in basic research and therapeutic applications. Several engineered Cas13 variants achieve targeted RNA degradation with minimal collateral effects.
 
Biomaterial properties of BPCDX
a, Appearance of BPCDX, indicating transparency and refractive nature of the curved device. b, Light transmission through 550-µm-thick samples of BPCDX, single-crosslinked BPC and the human cornea. The human cornea contains a layer of epithelial cells which absorb UV light⁶¹, whereas the bioengineered materials are cell-free. Data shown represent mean and standard deviation of measurements from three independent samples. c, Mechanical properties of BPCDX relative to single-crosslinked BPC and previously published data of bioengineered constructs made from porcine collagen29,30, with human cornea reference values³⁰ included for comparison. Data values for BPCDX represent mean and standard deviation of measurements from 22 independent samples per test (taken across different production batches, 550-µm-thick ‘dog-bone’ specimens). d, Scanning electron microscope images of the surface and bulk (cross-section) structure of BPCDX and a porcine cornea, indicating tightly packed collagen fibrils in BPCDX with diameter slightly thicker than the native porcine cornea (representative images from three samples per cornea type with similar results). e, Degradation of BPCDX, single-crosslinked BPC and a human donor cornea in 1 mg ml⁻¹ collagenase (data represent mean and standard deviation of measurements from three independent samples for bioengineered materials (550 µm thick, 12 mm diameter) and two independent samples of human donor cornea). f, HCE-2 human corneal epithelial cell attachment and growth on BPCDX relative to the control culture plate surface after 16 days of culture. Cells adhered to BPCDX, with NucBlue staining indicating nuclei and morphology of live, viable cells in brightfield mode. BPCDX had greater cell density than cell culture plasticware (three control samples, six BPCDX samples; error bars represent mean and standard deviation, P = 0.003, two-sided independent t-test). Scale bars, 100 µm.
Source Data
Results 6 months after intrastromal BPCDX implantation in minipigs
a, OCT images of the central 6 mm of cornea and corresponding photographs of operated eyes indicated localized thinning and loss of transparency in the access cut region in both groups (arrows in OCT scans and photographs). In the autograft group, one cornea (central image) exhibited loss of transparency, while another had partial loss of transparency (bottom image). In both cases, the implantation zone was skewed towards the limbus. The remaining three eyes were transparent with good thickness and only minor thinning at the access cut. In the BPCDX group, two eyes (second image from top, and bottom image) had a partially reduced transparency. In both cases, the implanted zone was skewed towards the limbus. In all eyes, transparency outside the access cut region was maintained. b, Pachymetry maps indicating corneal thickness 6 months post-operatively with color coding of thickness indicated by the grading scale, and mean thickness in µm indicated in each sector. BPCDX corneas exhibited similar thickness as the native porcine cornea. The native porcine cornea is shown for comparison purposes at the bottom of a,b. c, Table indicating pre-operative and post-operative mean corneal thickness and difference in corneal thickness in the central 2 mm zone as determined by OCT. Absence of positive fluorescein staining indicates complete epithelial wound healing. d, In vivo confocal microscopy images of porcine corneas at 6 months. In both groups, the epithelial cell mosaic appeared intact. Basal epithelium and sub-basal nerves (white arrows) were also observed, indicating preservation of the nerve plexus owing to minimal trauma during surgery. Anterior stromal nerves and keratocytes had normal morphology in both groups. The mid-stromal region appeared normal in autografts but BPCDX was devoid of keratocytes, except for individual cellular features (black arrows). Posterior stromal keratocytes and endothelial cells appeared normal and intact in both groups. All images in d are 400 × 400 µm². Representative images are from five corneas per group with similar results obtained for each group.
Source Data
Postmortem histologic analysis of corneas in the minipig model
a, Hematoxylin and eosin (H&E) staining revealed autografts with thickened epithelium and increased presence of anterior stromal cells relative to the native porcine cornea. Epithelial and stromal layers in BPCDX corneas were uniform, with maintenance of overall corneal structure and anatomy. Representative images shown from three corneas per group. b, Three different BPCDX-implanted corneas where host cells (arrows, left and center images) migrated into the BPCDX. The edge of the BPCDX had multiple tissue attachments (arrows, right image) with cells appearing to migrate towards the BPCDX. c, Immunohistochemical analysis indicated sub-basal nerves by the β-III-tubulin marker (immediately below the epithelium, arrows), while a preserved stromal nerve in a BPCDX cornea was apparent. Leukocyte marker CD45 indicated weak staining of stromal cells located at the BPCDX periphery (arrows), suggesting leukocyte-mediated remodeling⁴³ that differed in extent in different corneas (sections from two different BPCDX corneas shown). Leukocytes were absent in native and allograft corneas. All markers are indicated by green fluorescence, while a blue DAPI counterstain indicates the presence of cell nuclei. Non-specific diffuse signal from the green channel indicated the implanted BPCDX (asterisk in all panels). All images are representative images chosen from three corneas per group. Scale bars, 100 µm (a,c) and 50 µm (b).
Source Data
Clinical data from subjects in Iran receiving BPCDX
a, Keratometric and corneal thickness maps from the same subject indicate the thin and steep pre-operative cornea that was substantially thickened and flattened after intrastromal implantation of a 440-µm-thick BPCDX. The corresponding OCT cross-section scans indicate corneal thickness and shape before and after BPCDX implantation, with anterior and posterior borders of the BPCDX indicated by white arrows. The subject had an initial BSCVA of 20/200 that improved to 20/50 at 24 months post-operatively. b,c, Photographs of eyes from two subjects with the BPCDX four months post-operatively, indicating maintenance of corneal transparency. d,e, In vivo confocal microscopy images obtained from a single subject confirming the presence of sub-basal nerves (d) and endothelial cell mosaic (c) 6 months post-operatively. Endothelial cell density was 2,222 ± 62 cells per mm² in the eye. As only this single subject was imaged by in vivo confocal microscopy, it is unknown if these images are representative. Images in d,e are 400 × 400 µm².
Clinical data from a subject in India receiving BPCDX
a, Slit-lamp photographs pre-operatively (left) and one day post-operative (right) with arrows indicating immediate change in thickness and curvature in the central cornea. b, OCT scans indicating sustained thickening and regularization of corneal curvature following implantation of 280-µm-thick BPCDX (anterior and posterior surfaces of BPCDX indicated by white arrows). c, Topographic maps (left, values given are keratometric power in diopters), anterior surface elevation maps (center, values are in µm displacement from a best-fit sphere) and OCT pachymetric maps (right, thickness in µm) from the same subject indicated substantial flattening of the steepest pre-operative central region (black arrow), and substantial increase in corneal thickness post-operatively. The subject initially had a best contact lens-corrected visual acuity (BCLVA) of 20/600. At 24 months BCLVA improved to 20/30.
Article
Visual impairment from corneal stromal disease affects millions worldwide. We describe a cell-free engineered corneal tissue, bioengineered porcine construct, double crosslinked (BPCDX) and a minimally invasive surgical method for its implantation. In a pilot feasibility study in India and Iran (clinicaltrials.gov no. NCT04653922), we implanted BPCDX in 20 advanced keratoconus subjects to reshape the native corneal stroma without removing existing tissue or using sutures. During 24 months of follow-up, no adverse event was observed. We document improvements in corneal thickness (mean increase of 209 ± 18 µm in India, 285 ± 99 µm in Iran), maximum keratometry (mean decrease of 13.9 ± 7.9 D in India and 11.2 ± 8.9 D in Iran) and visual acuity (to a mean contact-lens-corrected acuity of 20/26 in India and spectacle-corrected acuity of 20/58 in Iran). Fourteen of 14 initially blind subjects had a final mean best-corrected vision (spectacle or contact lens) of 20/36 and restored tolerance to contact lens wear. This work demonstrates restoration of vision using an approach that is potentially equally effective, safer, simpler and more broadly available than donor cornea transplantation.
 
Article
mRNA printers will bring low-cost vaccines and made-to-order treatments for a range of different diseases.
 
Article
Hackathons not only advance science itself but also help to educate young biotechnologists and artificial intelligence enthusiasts and to build capacity in digital biotechnology in low-income settings.
 
Article
The elusive HER3, known to drive aggressive cancers, may finally be toppled by once-abandoned antibody drugs.
 
Article
A quarterly snapshot of job expansions, reductions and availability in the biotech and pharma sectors.
 
Article
Patients need the patent system to spur the investment necessary for the continued development of life-saving diagnostics and therapies.
 
Article
Improving target selectivity and defining cell phenotypes may improve the odds of finding therapies for fibrosis.
 
Article
Transplantation of B cells engineered ex vivo to secrete broadly neutralizing antibodies (bNAbs) has shown efficacy in disease models. However, clinical translation of this approach would require specialized medical centers, technically demanding protocols and major histocompatibility complex compatibility of donor cells and recipients. Here we report in vivo B cell engineering using two adeno-associated viral vectors, with one coding for Staphylococcus aureus Cas9 (saCas9) and the other for 3BNC117, an anti-HIV bNAb. After intravenously injecting the vectors into mice, we observe successful editing of B cells leading to memory retention and bNAb secretion at neutralizing titers of up to 6.8 µg ml⁻¹. We observed minimal clustered regularly interspaced palindromic repeats (CRISPR)–Cas9 off-target cleavage as detected by unbiased CHANGE-sequencing analysis, whereas on-target cleavage in undesired tissues is reduced by expressing saCas9 from a B cell-specific promoter. In vivo B cell engineering to express therapeutic antibodies is a safe, potent and scalable method, which may be applicable not only to infectious diseases but also in the treatment of noncommunicable conditions, such as cancer and autoimmune disease.
 
DVP concept and workflow
DVP combines high-resolution imaging, AI-guided image analysis for single-cell classification and isolation with an ultra-sensitive proteomics workflow². DVP links data-rich imaging of cell culture or archived patient biobank tissues with deep-learning-based cell segmentation and machine-learning-based identification of cell types and states. (Un)supervised AI-classified cellular or subcellular objects of interest undergo automated LMD and MS-based proteomic profiling. Subsequent bioinformatics data analysis enables data mining to discover protein signatures, providing molecular insights into proteome variation in health and disease states at the level of single cells. tSNE, t-distributed stochastic neighbor embedding.
BIAS for integrative image analysis and automated LMD single-cell isolation
a, AI-driven nucleus and cytoplasm segmentation of normal-appearing and cancer cells and tissue using BIAS. b, We benchmarked the accuracy of its segmentation approach using the F1 metric and compared results to three additional methods—M1 is unet4nuclei⁶, M2 is CellProfiler⁸ and M3 is Cellpose⁷—while OUR refers to nucleAIzer³. Bars show mean F1 scores with s.e.m.; n = 10 independent images for melanoma tissue and (U2OS) cells, and n = 20 for salivary gland tissue. Visual representation of the segmentation results: green areas correspond to true positive, blue to false positive and red to false negative. c, BIAS serves as the interface between the scanning and an LMD microscope, allowing high-accuracy transfers of cell contours between the microscopes. Illustration of cutting offset with respect to the object of interest and optimal path finding. d, Practical illustration of the functions in the upper panel. e, Immunofluorescence staining of the human fallopian tube epithelium with FOXJ1 and EpCAM antibodies, detecting ciliated and epithelial cells, respectively. Left panel: Ciliated (FOXJ1-positive) and secretory (FOXJ1-negative) cells. Right panel: Cell classification based on FOXJ1 intensity. Class 1 (FOXJ1-positive) and class 2 (FOXJ1-negative); magnification factor = ×387. f, PCA of FOXJ1-positive and FOXJ1-negative cell proteomes. g, Heat map of known protein markers for secretory and ciliated cells. Protein levels are z-scored. Asterisks represent imputed data. The marker list was derived from the Human Protein Atlas²⁰ project and based on literature mining. h, Volcano plot of the pairwise proteomic comparison between FOXJ1-positive and FOXJ1-negative cells. Cell-type-specific marker proteins are highlighted in green and turquoise, and black represents potential novel marker proteins. Significant enriched cell-type-specific proteins are displayed above the black lines (two-sided t-test, FDR < 0.05, s0 = 0.1, n = 4 biological replicates).
DVP defines single-cell heterogeneity at the subcellular level
a, Segmentation of whole cells and nuclei in BIAS of DNA (DAPI)-stained U2OS cells. Scale bar, 20 μm b, Automated LMD of whole cells and nuclei into 384-well plates. Images show wells after collection. c, Relative protein levels (x axis) of major cellular compartments between whole cell (n = 3 biological replicates) and nuclei (n = 3 biological replicates) specific proteomes. y axis displays point density. d, Left: conceptual workflows of the phenotype finder model of BIAS for ML-based classification of cellular phenotypes. Right: results of unsupervised ML-based classification of six distinct U2OS nuclei classes based on morphological features and DNA staining intensity. Colors represent classes. Scale bar, 20 μm. e, Phenotypic features used by ML to define six distinct nuclei classes. Radar plots show z-scored relative levels of morphological features (nuclear area, perimeter, solidity and form factor) and DNA staining intensity (total DAPI signal). f, Example images of nuclei from the six classes identified by ML. Blue color shows DNA staining intensity, and red color shows EdU staining intensity to identify cells undergoing replication. Represented nuclei are enlarged for visualization and do not reflect actual sizes. g, PCA of five interphase classes based on 3,653 protein groups after data filtering. Replicates of classes (n = 3 biological replicates) are highlighted by ellipses with a 95% confidence interval. h, Enrichment analysis of proteins regulated among the five nuclei classes. Significant proteins (515 ANOVA significant, FDR < 0.05, s0 = 0.1) were compared to the set of unchanged proteins based on Gene Ontology Biological Process (GOBP), Reactome pathways as well as cell cycle and cancer annotations derived from the Human Protein Atlas (HPA)²⁰. A Fisher’s exact test with a Benjamini–Hochberg FDR of 0.05 was used (Supplementary Table 3). i, Unsupervised hierarchical clustering of all 515 ANOVA significant protein groups (Supplementary Table 4). Cell-cycle-regulated proteins reported by the HPA are shown in the lower bar. Nuclei classes (n = 3 biological replicates) are shown in the row bar. C1–C4 show clusters upregulated in the different nucleus classes. j, Network analysis of enriched pathways for protein clusters C1–C4. Pathway enrichment analysis was performed with the ClusterProfiler R package³⁶. ER, endoplasmic reticulum; PC, principal component.
DVP applied to archived tissue of a rare salivary gland carcinoma
a, IHC staining of an acinic cell carcinoma of the salivary gland using the cell adhesion protein EpCAM. b, Representative regions from normal-appearing tissue (upper panels I and II) and acinic cell carcinoma (lower panels III and IV) from a. c, DVP workflow applied to the acinic cell carcinoma tissue. DL-based single cell detection of normal-appearing (green) and neoplastic (magenta) cells positive for EpCAM. Cell classification based on phenotypic features (form factor, area, solidity, perimeter and EpCAM intensity). d, Proteome correlations of replicates from normal-appearing (normal, n = 6) or cancer regions (cancer, n = 9). e, Volcano plot of pairwise proteomic comparison between normal and cancer tissue. t-test significant proteins (two-sided t-test, FDR < 0.05, s0 = 0.1, n = 6 biological replicates for normal and n = 9 for cancer) are highlighted by black lines. Proteins more highly expressed in normal tissue are highlighted in green on the volcanoʼs left, including known acinic cell markers (AMY1A, CA6 and PIP). Proteins more highly expressed in the acinic cell carcinoma are on the right in magenta, including the proto-oncogene SRC and interferon response proteins (MX1 and HLA-A; Supplementary Table 6). f, IHC validation of proteomic results. CNN1, SRC, CK5 and FASN are significantly enriched in normal or cancer tissue. Scale bar, 100 μm.
DVP applied to archived primary melanoma tissue
a, DVP sample isolation workflow to profile primary melanoma. b, DVP applied to primary melanoma immunohistochemically stained for the melanocyte marker SOX10 and the melanoma marker CD146. Left panel: stained melanoma tissue on a PEN glass membrane slide. Right panel: pathology-guided annotation of different tissue regions. Scale bar, 1 mm. c, Pathologist-guided and ML-based cell classification based on CD146 and SOX10 staining intensity and spatial localization: normal melanocytes, stromal cells, melanoma in situ, CD146-low melanoma, CD146-high melanoma, radial growth melanoma and vertical growth melanoma. Right lower panel: frequency of classes predicted by unsupervised ML (k-means clustering). d, Example pictures of the seven identified classes. Magnification factor = ×4,400. e, Correlation matrix (Pearson r) of all 27 measured proteome samples. f, PCA of proteomes. g, PCA of all melanoma-specific proteomes from in situ to invasive (vertical growth) melanoma. h, Unsupervised hierarchical clustering based on all 1,910 ANOVA significant (FDR < 0.05) protein groups. Two clusters of upregulated (cluster A) or downregulated (cluster B) proteins in invasive melanoma are highlighted. i, Tissue heat map mapping the proteomics results onto the imaging data. Relative pathway levels of selected terms from the two clusters are highlighted in i. Median protein levels were calculated per annotation and plotted for each isolated cell class against their x and y coordinates, as defined by their segmented cellular contours. j, Box plots of z-scored protein levels for the differentially regulated pathways visualized in i above. The box plots define the range of the data (whiskers), 25th and 75th percentiles (box) and medians (solid line). Outliers are plotted as individual dots outside the whiskers. k, Comparing proteomic changes in CD146-high melanoma cells (class 4) of the vertical growth (region 2) with the radial growth (region 1). Blood vessels in proximity to melanoma cells of the vertical growth are highlighted in red. Scale bar, 1 mm. l, Gene set enrichment analysis plot of significantly enriched pathways for melanoma cells of the vertical and radial growth phase. Pathway enrichment analysis was based on the protein fold change between vertical and radial melanoma cells and performed with the ClusterProfiler R package³⁶. Enriched terms with an FDR < 0.05 are shown. MHC, major histocompatibility complex.
Article
Despite the availabilty of imaging-based and mass-spectrometry-based methods for spatial proteomics, a key challenge remains connecting images with single-cell-resolution protein abundance measurements. Here, we introduce Deep Visual Proteomics (DVP), which combines artificial-intelligence-driven image analysis of cellular phenotypes with automated single-cell or single-nucleus laser microdissection and ultra-high-sensitivity mass spectrometry. DVP links protein abundance to complex cellular or subcellular phenotypes while preserving spatial context. By individually excising nuclei from cell culture, we classified distinct cell states with proteomic profiles defined by known and uncharacterized proteins. In an archived primary melanoma tissue, DVP identified spatially resolved proteome changes as normal melanocytes transition to fully invasive melanoma, revealing pathways that change in a spatial manner as cancer progresses, such as mRNA splicing dysregulation in metastatic vertical growth that coincides with reduced interferon signaling and antigen presentation. The ability of DVP to retain precise spatial proteomic information in the tissue context has implications for the molecular profiling of clinical samples.
 
Inhibition of RNR increased the efficiency of gene targeting in human and mouse cell lines
a, The GAPDH genomic locus after on-target HR with the rAAVDJ-GAPDH-P2A-GFP gene-targeting vector. The positions of in and out PCR primers, to detect site-specific integration, are indicated by black arrows. b, Huh7 cells were treated with RNR inhibitors and transduced with the gene-targeting vector rAAVDJ-GAPDH-P2A-GFP. Flow cytometry analysis of GFP⁺ fractions was performed after 14 d; Flu, fludarabine; Tri, triapine; Gal, gallium nitrate. Bars represent the group mean, and error bars represent s.d.; n = 3 biological replicate wells. Significance was determined by one-way analysis of variance (ANOVA) with Dunnett’s test for multiple comparisons. P values of all groups were <0.0001, except Gal, which was 0.9998. c, Huh7 cells were treated with RNR inhibitors and transduced with an rAAVDJ vector expressing Firefly luciferase (Fluc) from the CAG promoter. Luciferase activity was measured 24 h later; RLU, relative light units. Data from b and c are representative of two independent experiments. Values are displayed as the group means, with error bars representing s.d.; n = 3 biological replicate wells. Significance was determined by one-way ANOVA analysis with Dunnett’s test for multiple comparisons. d,e, Huh7 cells were transfected with RRM1 small interfering RNA (siRNA; siRRM1) and transduced with rAAVDJ-GAPDH-P2A-GFP; siScr, scrambled siRNA control. Western blotting of RRM1 was performed 2 d after siRNA transfection (d). The GFP⁺ fraction was analyzed by flow cytometry 3 d after transduction (e). Bars represent the group mean, and error bars represent s.d.; n = 3 biological replicate wells. Data are representative of two independent experiments. Significance was determined using a two-tailed t-test. f, A junction capture PCR was used to detect on-target integration at the GAPDH locus using gDNA extracted from Huh7 cells treated with the indicated RNR inhibitors 14 d after AAV transduction. Control reactions included amplification of the ACTB locus; M, size marker. g, Murine Hepa1-6 cells were treated with RNR inhibitors and transduced with the rAAVDJ-Alb-P2A-GFP targeting vector. Flow cytometry analysis of GFP⁺ fractions at 14 d after AAV transduction is shown. Bars represent the group mean, and error bars represent s.d.; n = 3 biological replicate wells. Data are from two independent experiments. The HU group P values were 0.0002 and 0.0076 for the fludarabine group. h, The mouse Alb locus after HR with the gene-targeting rAAVDJ-Alb-P2A-GFP vector is shown. The positions of quantitative PCR (qPCR) primers to detect on-target integrated fusion mRNA are indicated; Ex, exon. i, RNA was extracted from transduced Hepa1-6 cells, and qPCR was performed to quantify expression levels of on-target Alb-P2A-GFP fusion mRNA. Actb mRNA was used for normalization, and data are shown as relative expression to control. Bars represent the group mean, and error bars represent s.d.; n = 3 biological replicate wells. Data are from two independent experiments. Significance was determined by one-way ANOVA with Dunnett’s test for multiple comparisons for all data, unless otherwise indicated; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not significant.
Source data
Fludarabine administration increased the efficiency of gene targeting in mouse hepatocytes
a, Four-week-old mice were administered intraperitoneal (i.p.) injections of HU (300 mg kg–1) once per day or fludarabine (125 mg kg–1) three times per day through day 1 to day 3. Mice were also administered intravenous (i.v.) injections with the rAAV8-Alb-P2A-hF9 targeting vector (1.0 × 10¹¹ viral genomes per mouse) on day 1 immediately after the HU or second fludarabine injection. b, hF9 protein levels in mouse plasma were determined using an enzyme-linked immunosorbent assay (ELISA) following rAAV8-Alb-P2A-hF9 targeting vector injection with or without drug treatment over a 65-d period. Values are displayed as the group mean with error bars representing s.d.; n = 5 mice per group. Significance testing was performed by two-way ANOVA analysis. c, gDNA was extracted from liver tissues 65 d after rAAV8-Alb-P2A-hF9 injection, and qPCR was performed to quantify total AAV genomes. Actb primers were used to quantify mouse diploid genomes. Each point represents data from one mouse. Bars represent the group mean, and error bars represent the s.d.; n = 5 mice per group. Significance testing was performed using a one-way ANOVA analysis followed by Dunnett’s t-test. d, The mouse Alb locus after HR with the gene-targeting AAV-Alb-P2A-hF9 vector is shown. Exon–intron structure and the positions of qPCR primer pairs used for e–g are indicated; HA, homology arm; Fw, forward; Rv, reverse. e–g, Total RNA was extracted from mouse liver tissues in Fig. 2b. qPCR assays quantified the expression levels of on-target integration-derived Alb-P2A-hF9 fusion mRNA (primers Fw1 and Rv2; arrows) (e), total hF9 mRNA (Fw2 and Rv3) (f) and the fraction of hF9 fusion mRNA derived from on-target HR out of the total amount of hF9 mRNA (g). Actb mRNA was used for normalization, and data are shown as relative expression to the PBS-treated group. Bars represent the group mean, and error bars represent the s.d.; n = 5 mice per group. Significance was determined using a one-way ANOVA analysis followed by Dunnett’s t-test, unless otherwise indicated. h, One-week-old male neonatal mice were injected i.p. with PBS or fludarabine (375 mg kg–1) and 4 h later with rAAV8-Alb-P2A-hF9 (2.5 × 10¹³ viral genomes per kilogram of body weight). A second fludarabine dose was given 1 d after vector injection. Four weeks later, plasma was drawn, and hF9 levels were measured by ELISA; n = 6 PBS-treated mice and n = 8 fludarabine-treated mice. Values are displayed as the group mean, with error bars representing s.d. Significance testing was performed by two-way ANOVA. i, Fludarabine dosing regimens were tested by i.p. administration (125 mg kg–1) three times per day for one, three or five sequential days. Four-week-old mice were injected i.v. at day 1 with the rAAV8-Alb-P2A-hF9 targeting vector (1.0 × 10¹¹ viral genomes per mouse) immediately after the second fludarabine administration injection. Blood was collected 2 months later, and hF9 protein levels were determined via ELISA. Values are displayed as the group mean with error bars representing s.d.; n = 4 mice per group (see also Extended Data Fig. 6). Significance was determined using a one-way ANOVA followed by Dunnett’s t-test; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Fludarabine transiently inhibits S phase progression, and rAAV gene targeting occurs in hepatocytes that have not progressed through S phase
a, Mice were injected i.p. with BrdU (200 mg kg–1) once per day for 3 d to label proliferating hepatocytes. Some mice were simultaneously injected with fludarabine (125 mg kg–1 three times per day for 3 d), while the final group was treated with fludarabine before the 3 d of BrdU injection; n = 3 mice per group. b, Six hours after the last injection, mice were killed, and livers were collected for immunostaining with anti-BrdU. Representative images from each group are shown with BrdU-labeled nuclei (red) and a DAPI counterstain (blue). All images were taken with a ×20 objective with identical exposure and settings. c, Images of BrdU-labeled nuclei were quantified from each group as the number of BrdU⁺ nuclei per field of view. Images used for quantification are from two or more slides per mouse, three mice per group and two or more independent strains. Significance was determined using a Shapiro–Wilk test for normal distribution and two-tailed Student’s t-test for normally distributed data or non-parametric Mann–Whitney U-test for non-normally distributed data. Each point represents average data from one mouse; n = 3 per group. Bars represent the group mean, and error bars represent the s.d. The P value of the fludarabine acute group is 0.0012, and the P value for the fludarabine washout group is 0.4. d, To label proliferating cells in the liver, mice were i.p. injected with BrdU once per day for 7 d. On day 1 of BrdU injections, two groups of mice were i.v. injected with the rAAV8-Alb-P2A-GFP targeting vector (1 × 10¹¹ viral genome per mouse). One group was also treated with fludarabine (125 mg kg–1, three times per day) for the first 3 d. e, On day 7, mice were killed, and livers were collected for immunohistochemistry staining of BrdU (red) and GFP (green) with a DAPI nuclear stain (blue). Representative images are shown. f, GFP⁺ cells were quantified as the number of positive cells per 20× field of view. An F-test was used to determine variance between groups of normally distributed data, and a two-tailed Student’s t-test or t-test with Welch’s correction was used to test for significance. Images used for quantification are from two or more slides per mouse with three mice per group and two or more independent strains. Each point represents average data from one mouse; n = 3 per group. Bars represent the group mean, and error bars represent s.d. g,h, Total RNA was extracted from liver tissue, and qPCR was used to determine the levels of on-target integration-derived Alb-P2A–GFP mRNA and endogenous Alb mRNA, as described previously. Each point represents average data from one mouse; n = 3 per group. Bars represent the group mean, and error bars represent s.d. Statistical testing was performed using a one-way ANOVA followed by Dunnett’s t-test; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. i, The number of GFP⁺ cells that colocalized with BrdU (that is, the number of cells that have undergone S phase DNA synthesis) were quantified and displayed as the percentage of total GFP⁺ cells per group.
Fludarabine induces a transient DNA damage response in mice
a, Liver tissue sections from the same mice in Fig. 3a–c were also stained for the DNA damage response marker γH2AX. Representative images are shown with γH2AX (red) and DAPI (blue). b, Images of γH2AX nuclei were quantified from each group and are displayed as the percentage of γH2AX⁺ nuclei out of all nuclei. Images used for quantification are from two or more slides per mouse, three mice per group and two or more independent strains. An F-test was used to determine variance between groups of normally distributed data, and a two-tailed Student’s t-test or t-test with Welch’s correction was used to test for significance. Each point represents average data from one mouse; n = 3 per group. Bars represent group means, and error bars represent s.d. c, Liver tissue lysates from the same mice were used for western blotting of γH2AX, and α-tubulin was used as a loading control (top). Image analysis quantification of the γH2AX band intensity in the western blot was normalized to α-tubulin (bottom). Each lane (top) and point (bottom) represents average data from one mouse; n = 3 mice per group, except the DEN data, which are two technical replicates from one mouse. Error bars represent s.d.
Fludarabine increases CRISPR/Cas9 gene editing efficiency in vivo
a, saCas9 is targeted to the intron downstream of exon 14 in the mouse Alb locus, delivered by AAV-saCas9. Homology-directed repair is accomplished using an AAV-encoded repair template, AAV-GFP-HDR, which contains a P2A-GFP transgene flanked by sequences homologous to the gRNA target site. The PAM site of AAV-GFP-HDR was mutated to avoid self-targeting; 3′-UTR, 3′-untranslated region. ITR, inverted terminal repeats. b, Mice were treated with PBS or fludarabine, as previously described, and co-injected with AAV-saCas9 and AAV-GFP-HDR. Three ratios of AAV-saCas9 to AAV-GFP-HDR vector were used (1:1, 1:5 and 1:10), with each part representing 6.0 × 10¹² viral genomes per kilogram of body weight. Two weeks later, mice were sacrificed, and livers were imaged for native GFP expression. Representative images are shown; n = 3 to 5 mice per group. c, GFP⁺ cells were quantified from multiple images per mouse. A Shapiro–Wilk test was used to test for normal distribution, an F-test determined variation between groups, and a two-tailed Student’s t-test or t-test with Welch’s correction tested for significance. Each point represents data from an individual mouse; n = 3 to 5 mice per group. Bars represent the group mean, with error bars representing s.d. The P value was 0.0012 for the 1:1 group, 0.0002 for the 1:5 group and 0.138 for the 1:10 group. d, Total RNA was extracted from mouse livers and used for quantification of the on-target HR-derived fusion mRNA or Alb mRNA by qPCR. mRNA levels were normalized to Actb mRNA. Data are displayed as a percentage of fusion mRNA out of all Alb mRNA. Each point represents data from an individual mouse; n = 3 to 5 mice per group. Bars represent the group mean, with error bars representing s.d. Significance testing was performed as in c. The P value was 0.0084 for the 1:1 group, 0.004 for the 1:5 group and 0.057 for the 1:10 group. e, Separately, mice were injected with 6.0 × 10¹² viral AAV-saCas9 genomes per kilgram of body weight with or without fludarabine treatment. Two weeks later, gDNA was extracted from livers, and targeted deep sequencing of the gRNA target site was performed. Data are normalized to sequencing data from a non-injected control mouse and are displayed as a percentage of alleles containing indels; n = 3 mice per group. Bars represent the group mean, and error bars represent s.d. Significance testing was performed with a two-tailed t-test after testing for distribution and variance. f, The top 12 mutant alleles from targeted deep sequencing were analyzed for all mice and technical replicates. Indels were graphed as the percentage of reads from all mutant alleles based on size (x axis).
Article
Homologous recombination (HR)-based gene therapy using adeno-associated viruses (AAV-HR) without nucleases has several advantages over classic gene therapy, especially the potential for permanent transgene expression. However, the low efficiency of AAV-HR remains a major limitation. Here, we tested a series of small-molecule compounds and found that ribonucleotide reductase (RNR) inhibitors substantially enhance AAV-HR efficiency in mouse and human liver cell lines approximately threefold. Short-term administration of the RNR inhibitor fludarabine increased the in vivo efficiency of both non-nuclease- and CRISPR/Cas9-mediated AAV-HR two- to sevenfold in the murine liver, without causing overt toxicity. Fludarabine administration induced transient DNA damage signaling in both proliferating and quiescent hepatocytes. Notably, the majority of AAV-HR events occurred in non-proliferating hepatocytes in both fludarabine-treated and control mice, suggesting that the induction of transient DNA repair signaling in non-dividing hepatocytes was responsible for enhancing AAV-HR efficiency in mice. These results suggest that use of a clinically approved RNR inhibitor can potentiate AAV-HR-based genome-editing therapeutics.
 
Highly diverse GF prediction with different tools
a, GFs (n = 2,361) were predicted with five prediction tools for the MCF7 breast cancer cell line from two sequencing replicates. b, The number of distinct predicted GFs (y axis) in MCF7 is shown for different combinations of tools (x axis) as indicated by dots below the bars. The fraction of GFs identified by one tool (orange) or multiple tools (green) is shown in the pie chart, and the percentage of GFs identified in two sequencing replicates (gray diagonal pattern) for one and multiple tool predictions is shown in the horizontal bar chart. c, 133 GFs were validated by RT–qPCR in MCF7 and SKBR3. Shown is the fraction of positive (blue) and negative (red) tested GFs according to the number of detecting tools and according to identification in one or both sequencing replicates. Labels at the top indicate number of tested GFs. d, Predicted GFs for 14 primary triple-negative breast cancer (TNBC) samples are shown (n = 4,488; one tool = orange, multiple tools = green). The pie chart shows the fraction of GFs predicted by one tool (orange) and multiple tools (green). e, 492 GFs were validated by RT–qPCR in 14 primary TNBC samples. Shown is the fraction of positive (blue) and negative (red) tested GFs according to the number of detecting tools. Labels at the top indicate the number of tested GFs. f, Considering the confirmation rate from RT–qPCR and the predicted number of GFs according to the number of predicting tools, the number of true-positive GFs was estimated (one tool = light blue, multiple tools = dark blue).
Recurrent GFs are enriched for cis-near fusions in normal tissue
a, The frequency of recurrence is shown for distinct GFs in 14 TNBC samples (bar chart). The proportion of recurrently (dark gray) and uniquely (light gray) predicted GFs is also depicted (pie chart). b, Recurrent GFs are enriched for cis-near configuration. The number of recurrent and unique GFs is shown according to the configuration type of the breakpoints. ‘cis’ indicates GFs with breakpoints on same chromosome, ‘trans’ on different chromosome. ‘inv’ indicates breakpoints on different strands. ‘cis-near’ indicates GFs with breakpoints on the same chromosome and strand, within 1 Mb distance, whereas ‘cis-far’ indicates GFs with breakpoints farther apart. Percentages and total numbers of GFs are indicated. c, The overlap between recurrent (dark gray) and unique (light gray) predicted GFs in 14 TNBC samples with four normal breast tissue samples (green) is shown (Venn). The proportion of GFs in ‘cis-near’ configuration (green) for shared recurrent and unique GFs is shown (pie chart).
EasyFuse improves sensitivity in detecting tumor-specific GFs
a, The EasyFuse pipeline consists of five major steps for GF prediction: (1) STAR-based read filtering excludes non-aberrant mapping reads to improve subsequent prediction performance; (2) execution of five evaluated prediction tools on filtered data; (3) uniform annotation of affected transcript variants and construction of context sequences across breakpoints with custom scripts; (4) quantification of supporting junction and spanning reads from STAR-mapped reads to context sequences; and (5) ranking of prediction candidates with a random forest machine learning classifier trained on RT–qPCR-validated GFs. b, The fraction of reads remaining after the read-filtering step is shown for 14 FF breast cancer samples. c, Tool runtimes are shown in minutes with and without read filtering derived for n = 14 FF breast cancer samples. The line indicates the median value; the box represents the interquartile range; and whiskers extend by 1.5-fold interquartile range. d,e, Changes in predicted (d) and RT–qPCR-confirmed (e) ‘cis-near’ and ‘trans-like’ GFs by read filtering. Negative numbers indicate loss in prediction (white); positive numbers indicate gain (purple); and unaltered predictions are in gray. The fraction of GFs identified by one (orange) or two (green) tools in filtered and unfiltered read data is shown (pie chart).
Machine learning contributes to highly specific prediction of GFs in FFPE tumor samples
a, The number of predicted GFs (as unique breakpoint pairs) for 14 tumor samples. Breakpoint pairs that were identified in both replicates are shown in dark gray, whereas those found in single replicates are in light gray. The tumor type, histological tumor content of the sample and whether it is a primary (P) or metastasis (M) sample is indicated below. b, A total of 853 fusion breakpoints detected by EasyFuse across all samples were validated by RT–qPCR (positive in blue and negative in red); validation data were separated by samples into training and test datasets. c, The number and percent of GF breakpoints per sample detected in a single replicate or both replicates (top) as well as the number and percent of fusions validated positive (blue) or negative (red) in these subsets. d, Features used in the machine learning models. The heat map indicates which features (rows) are used in the different models (columns). The relative feature importance in the random forest model is shown. ft refers to the GF, and wt1 and wt2 refer to the corresponding wild-type variants. A more detailed description of all features can be found in Supplementary Table 10. e,f, The performance of four different random forest models on confirmation data of 281 candidate GF calls from EasyFuse in replicates from the test samples that were not used for model training are shown as receiver operating characteristic curves (e) and precision–recall curves (f). The models were trained on different subsets of features as indicated in d. g, Benchmark of fusion prediction performance between EasyFuse with the model ‘EF_full’ and six other tools on all validated fusions genes in the three test samples. PPV (top), sensitivity (middle) and F1 score (bottom) were calculated by weighting according to the concordance between tools as described in the Methods. h, Performance separately for GFs in cis-near (same chromosome and strand, less than 1 Mb distance) and trans-like configuration. The concordance bins were recalculated on the two subsets of fusions for weighted performance calculation.
Predicted GFs encode immunogenic neo-antigens eliciting CD4⁺ and CD8⁺ T cell responses
a, The EasyFuse machine learning model predicted a median of 46 GFs in 14 patients with melanoma. b, Candidate filtering and epitope prediction resulted in a median of nine GFs per patient that encode at least one MHC class I or class II epitope. c, IFN-γ ELISpots after IVS were carried out for 30 selected fusion peptides with patient-derived PBMCs and resulted in CD4⁺ T cell responses for ten fusion peptides. CD4⁺ T cell reactivity against GF targets for PA-043, PA-045, PA-046, uID004 and uID018 (samples indicated by *) were not determined owing to high ELISpot background. d, Quantification of IFN-γ CD4⁺ T cell responses after IVS (11 days) and re-stimulation with iDCs pulsed with target pool peptides for the respective patient sample. iDCs loaded with irrelevant peptides served as control. Data are shown as background-corrected mean spot count. e, Post-IVS-measured CD4⁺ and CD8⁺ T cell response to target peptide-loaded iDCs exemplified for one patient (^ indicates predicted breakpoint position). f, Predicted binding affinity of GF-encoded neo-epitopes, which were tested for CD8⁺ T cell response (MHC class I) or CD4⁺ T cell response (MHC class II). Binding affinity is shown in nanomolar (nM) as predicted by netMHCpan for MHC class I epitopes and netMHCIIpan for MHC class II epitopes with patient-specific alleles. The dotted line indicates an affinity of 500 nM. Peptides without predicted epitopes are not shown. g, To establish the expected GF-derived neo-antigens in breast cancer, we analyzed 57 TNBC samples with EasyFuse. A median of 12 neo-antigens with at least one predicted MHC class I or class II epitope was detected per sample. From all targets, 5% were found in more than one sample (inlay).
Source data
Article
Cancer-associated gene fusions are a potential source for highly immunogenic neoantigens, but the lack of computational tools for accurate, sensitive identification of personal gene fusions has limited their targeting in personalized cancer immunotherapy. Here we present EasyFuse, a machine learning computational pipeline for detecting cancer-specific gene fusions in transcriptome data obtained from human cancer samples. EasyFuse predicts personal gene fusions with high precision and sensitivity, outperforming previously described tools. By testing immunogenicity with autologous blood lymphocytes from patients with cancer, we detected pre-established CD4+ and CD8+ T cell responses for 10 of 21 (48%) and for 1 of 30 (3%) identified gene fusions, respectively. The high frequency of T cell responses detected in patients with cancer supports the relevance of individual gene fusions as neoantigens that might be targeted in personalized immunotherapies, especially for tumors with low mutation burden. EasyFuse detects gene fusions in cancer transcriptomes for personalized immunotherapy.
 
Article
Spatial transcriptomics enables the simultaneous measurement of morphological features and transcriptional profiles of the same cells or regions in tissues. Here we present multi-modal structured embedding (MUSE), an approach to characterize cells and tissue regions by integrating morphological and spatially resolved transcriptional data. We demonstrate that MUSE can discover tissue subpopulations missed by either modality as well as compensate for modality-specific noise. We apply MUSE to diverse datasets containing spatial transcriptomics (seqFISH+, STARmap or Visium) and imaging (hematoxylin and eosin or fluorescence microscopy) modalities. MUSE identified biologically meaningful tissue subpopulations and stereotyped spatial patterning in healthy brain cortex and intestinal tissues. In diseased tissues, MUSE revealed gene biomarkers for proximity to tumor region and heterogeneity of amyloid precursor protein processing across Alzheimer brain regions. MUSE enables the integration of multi-modal data to provide insights into the states, functions and organization of cells in complex biological tissues. Deep learning analysis of microscopy and spatial transcriptomics data characterizes cell types and states.
 
ScCUT&Tag-pro enables simultaneous profiling of CUT&Tag and protein levels
a, Schematic of experimental workflow, which is compatible with the 10x Chromium system. b, Distribution of unique fragments obtained per cell for six histone modifications, profiled in separate PBMC scCUT&Tag-pro experiments (left to right: n = 12,770, 9,575, 10,386, 8,304, 15,609 and 8,232 cells). The center, bounds and whiskers of the boxplot show median, quartiles and data points that lie within 1.5× interquartile range of the lower and upper quartiles, respectively. Data beyond the end of the whiskers are plotted individually. c, Pseudobulk profiles of scCUT&Tag are well correlated with bulk CUT&Tag profiles of human PBMCs. d, Tornado plots of genomic regions ordered by chromatin accessibility. We observe no enrichment of H3K27me3 in accessible regions, indicating minimal open chromatin bias. e, Relationship between the number of cells included in a pseudobulk CUT&Tag profile of human PBMCs and the Pearson correlation with a bulk experiment. f, Uniform manifold approximation and projection (UMAP) visualizations of 12,770 single cells profiled with H3K4me1 scCUT&Tag-pro and clustered on the basis of CUT&Tag profiles, cell surface protein levels and WNN analysis, which combines both modalities. Cluster labels are derived from WNN analysis. g, Comparing pseudobulk CUT&Tag-pro profiles with ChIP-seq data from ENCODE. Including all cells assigned to each cell type results in pseudobulk tracks that closely mirror ENCODE profiles. However, even when downsampling to 300 cells per cluster, cell-type-specific patterns can still be observed. Ab, antibody; CTL, cytotoxic T lymphocyte; DC, dendritic cell; GCLC, glutamate-cysteine ligase catalytic subunit; MAIT, mucosal-associated invariant T cell; Mono, monocyte; NK, natural killer; pDC, plasmacytoid dendritic cell.
Protein measurements facilitate integrated analysis across modalities
a, Schematic workflow for integrated analysis. Datasets produced by scCUT&Tag-pro, ASAP-seq and CITE-seq are integrated together on the basis of a shared panel of cell surface protein measurements. b, Left: UMAP visualization of 230,597 total cells projected onto the reference dataset from Hao et al.³⁰. Right: In addition to a harmonized visualization, cells from all experiments are annotated with a unified set of labels. c, We learned a unified pseudotime trajectory based on all experiments, representing the CD8 T cell transition from naive to effector states. We observe identical molecular dynamics for naive (CD55), memory (CD103) and effector (CD57) markers across all experiments, demonstrating that integrative analysis accurately identifies cells in matched biological states across experiments. d, Visualization of nine molecular modalities at the CD8A locus in B cell, CD4 T cell, CD8 T cell, dendritic cell, monocyte and NK cell groups. ASDC, AXL⁺ dendridic cell; cDC2, conventional dendridic cell 2; dnT, double-negative T cell; Eryth, erythrocyte; gdT, γδ T cell; HSPC, human stem and progenitor cell; TCM, central memory T cell; TEM, effector memory T cell; Treg, regulatory T cell.
ScChromHMM annotates chromatin states at single-cell resolution
a, Chromatin states returned by ChromHMM, which was run on 25 pseudobulk tracks for six histone marks. States are broadly grouped into five categories. b, Correlation comparing cell-type-specific pseudobulk profiles of H3K4me1 in CD14 monocytes generated from the original experiment or the interpolated values. Each point corresponds to a 200-bp genomic window (Methods) (c), For each cell type, the interpolated and original profiles are highly correlated and clustered together. d, scChromHMM outputs at the PAX5 locus. Top: Pseudobulk profiles for six chromatin marks in three cell types. Yellow bar highlights a 200-bp genomic window near the TSS. Bottom: scChromHMM posterior probabilities representing the annotation for the highlighted window in each cell. The region is uniformly annotated with a promoter state in B cells, where PAX5 is transcriptionally active, and as a repressive state in other cell types. e, Metaplots exhibiting the enrichment of chromatin accessibility and histone modifications at functional regions identified by ChromHMM (left) and scChromHMM (right) in CD14 monocytes. K, kilobase.
Extensive heterogeneity in repressive chromatin encodes cellular identity
a, Remodeling of repressive chromatin during CD8 T cell maturation. Heatmap shows the posterior probabilities (repressive state) in single cells for 14,585 genomic loci, as returned by scChromHMM. Cells are ordered by their progression along pseudotime (Fig. 2c). b, chromVAR deviation scores for the TBX21 and LEF1 motifs in single cells, ordered by their progression along pseudotime. We used the scChromHMM-derived posterior probabilities as input to chromVAR instead of chromatin accessibility levels. c, Unsupervised analysis of scChromHMM-derived probabilities (repressive state) separates granular cell types. d, Single-cell correlation matrix based on repressive chromatin at TSSs (Methods) when using all TSSs (left heatmap) or after excluding the top 3,000 transcriptionally variable genes (right heatmap). In each case, the observed correlation structure is fully consistent with cell type labels, suggesting that there is extensive heterogeneity in repressive chromatin even for genes that do not vary transcriptionally. e, Scatter plot showing average gene expression levels for all genes in CD14 monocytes (x axis) and other cell types (y axis). Colored points represent 1,597 loci where we detect changes in repressive chromatin for monocytes (Methods). Blue points represent 1,340 loci where we do not detect an accompanying transcriptional change. Red points represent 257 genes where we detect a transcriptional shift. TPM, transcripts per kilobase million. f, Four representative examples of individual genes shown as blue points in e.
Supervised mapping of scCUT&Tag datasets
a, UMAP visualization of 8,362 H3K27me3 scCUT&Tag profiles of human PBMCs from Wu et al.¹⁹ based on an unsupervised analysis and clustering. b, Same cells as in a, but after mapping to the multimodal reference defined in this paper. Cells are colored by their reference-derived level 2 annotations. c–e, Coverage plots showing the cell-type-specific binding patterns of H3K27me3 at three loci, MYO1C (c), ALOX5AP (d) and HSPA12A (e). Plots are shown for our dataset (reference), as well as the scCUT&Tag profiles from the query dataset (query, Wu et al.¹⁹). Cells in the query dataset are grouped by their predicted labels. We observe highly concordant patterns across datasets for all loci, supporting the accuracy of our predictions. Four representative cell types are shown at each locus. TEM, T effector memory.
Article
Technologies that profile chromatin modifications at single-cell resolution offer enormous promise for functional genomic characterization, but the sparsity of the measurements and integrating multiple binding maps represent substantial challenges. Here we introduce single-cell (sc)CUT&Tag-pro, a multimodal assay for profiling protein–DNA interactions coupled with the abundance of surface proteins in single cells. In addition, we introduce single-cell ChromHMM, which integrates data from multiple experiments to infer and annotate chromatin states based on combinatorial histone modification patterns. We apply these tools to perform an integrated analysis across nine different molecular modalities in circulating human immune cells. We demonstrate how these two approaches can characterize dynamic changes in the function of individual genomic elements across both discrete cell states and continuous developmental trajectories, nominate associated motifs and regulators that establish chromatin states and identify extensive and cell-type-specific regulatory priming. Finally, we demonstrate how our integrated reference can serve as a scaffold to map and improve the interpretation of additional scCUT&Tag datasets.
 
Article
Despite their clinical success, chimeric antigen receptor (CAR)-T cell therapies for B cell malignancies are limited by lengthy, costly and labor-intensive ex vivo manufacturing procedures that might lead to cell products with heterogeneous composition. Here we describe an implantable Multifunctional Alginate Scaffold for T Cell Engineering and Release (MASTER) that streamlines in vivo CAR-T cell manufacturing and reduces processing time to a single day. When seeded with human peripheral blood mononuclear cells and CD19-encoding retroviral particles, MASTER provides the appropriate interface for viral vector-mediated gene transfer and, after subcutaneous implantation, mediates the release of functional CAR-T cells in mice. We further demonstrate that in vivo-generated CAR-T cells enter the bloodstream and control distal tumor growth in a mouse xenograft model of lymphoma, showing greater persistence than conventional CAR-T cells. MASTER promises to transform CAR-T cell therapy by fast-tracking manufacture and potentially reducing the complexity and resources needed for provision of this type of therapy.
 
Overview of the CellTrek workflow
CellTrek first coembeds scRNA-seq and ST datasets into a shared latent space. Using the ST data, CellTrek trains a multivariate RF model with spatial coordinates as the outcome and latent features as the predictors. A 2D spatial interpolation on the ST data is introduced to augment the ST spots. The trained RF model is then applied to the coembedded data (ST interpolated) to derive an RF distance matrix, which will be converted into a sparse graph using MNN. Based on the sparse graph, CellTrek transfers the coordinates to single cells from their neighboring ST spots.
CellTrek reconstructs spatial organization in a mouse brain tissue
a, Comparison of CellTrek, NVSP-CellTrek and SrtCT results for single-cell spatial charting in a mouse brain tissue. b, KL-divergence of spatial cell charting methods for each cell type using SrtLT as a reference. c, UMAP (left) and CellTrek map (right) of scRNA-seq data of L5 IT cell states. d, Spatial colocalization graph of glutamatergic neurons using SColoc. e, CellTrek-based spatial K-distance of glutamatergic neurons to L2/3 IT cells (L2/3 IT n = 739; L4 n = 576; L5 IT n = 690; L5 PT n = 365; NP n = 182; L6 IT n = 695; L6 CT n = 721; L6b n = 153). Boxplots show the median with interquartile ranges (25–75%); whiskers extend to 1.5× the interquartile range from the box. We performed a Spearman correlation test. f, Spatial coexpression modules (K1 and K2) identified in L5 IT cells using SCoexp. g–h, UMAPs of L5 IT cells showing the K1 module activity scores (g) and the K2 module activity scores (h) and their corresponding CellTrek maps.
CellTrek reconstructs spatial organization in a mouse kidney tissue
a, Comparison of CellTrek, NVSP-CellTrek and SrtCT results for single-cell spatial charting in a mouse kidney tissue. DistTub, distal tubule cells; T, T cells; ProxTub, proximal tubule cells; VSMC, vascular smooth muscle cells; Inter, intercalated cells; Prin, principal cells; TLLH, the loop of Henle; Vasc, vascular cells; Macro, macrophages; RenaCorp, renal corpuscle cells. b, KL-divergence of spatial cell charting methods for each cell type using SrtLT as a reference. c, Trajectory analysis for proximal tubule cells (left) and spatial mapping of the pseudotime values in the tissue section (right). d, Trajectory analysis for distal tubule cells (left) and spatial mapping of the pseudotime values in the tissue section (right). e, Spatial colocalization graph of different renal cell types using SColoc. f, Spatial consensus matrix of different renal cell types. g, CellTrek-based spatial K-distance of TLLH, DistTub and Prin cells to the tissue center cells across experimental zonal dissections (left). Center cells as reference are shown on the right panel. Kruskal–Wallis rank sum tests were performed (cortex n = 1,169; outer medulla n = 1,423; inner medulla n = 1,661). Boxplots show the median with interquartile ranges (25–75%); whiskers extend to 1.5× the interquartile range from the box. h, Spatial coexpression modules (K1 and K2) identified in distal tubule cells using SCoexp. i–j, UMAPs of distal tubule cells showing the K1 module activity scores (i) and the K2 module activity scores (j) and their corresponding CellTrek maps.
CellTrek identifies the spatial subclone heterogeneity in DCIS1
a, A heatmap of copy number (CN) profiles inferred by CopyKAT on the scRNA-seq data in DCIS1. The lower part represents a consensus CN profile of each cluster with some breast cancer-related genes annotated. b, CN-based UMAP of DCIS1. c, Phylogenetic tree based on the consensus CN profiles. d, Hallmark GSEA analysis of the expression data from three tumor subclones. P values were determined by an adaptive multi-level split Monte-Carlo method and then adjusted by the Benjamini-Hochberg method. e, Spatial cell charting of three tumor subclones using CellTrek. f, Tumor subclonal compositions in different ducts. The diamond symbol in each bar represents the Shannon index which measures the diversity of tumor subclones. g, H&E image of the DCIS tissue section with Shannon diversity index for each duct.
CellTrek displays the spatial tumor-immune microenvironment in DCIS2
a, H&E image of the tissue section from the DCIS2 patient. Histopathological annotations of tumor regions are highlighted in red circles with labels from T1 to T11. b, UMAP of DCIS2 scRNA-seq data (tumor cells, B cells, NK/T cells, myeloid and pDC cells). c, CellTrek spatial mapping of tumor cells, B cells, NK/T cells, myeloid and pDC cells. Yellow boxes highlight potential locations of tertiary lymphoid structures (TLS) with aggregation of mixed immune cells. d, ST spot-level TLS signature scores. e, Boxplot showing the association between CellTrek-based immune cell counts and ST spots in TLS score quantiles (each quantile contains 61-62 ST spots). We performed a Spearman correlation test. f, CellTrek spatial mapping of different T cell states. The contour plot represents the tumor cell densities. g, UMAP of scRNA-seq data showing different T cell states. h, Spatial colocalization graph of T cell states using SColoc. i, CellTrek spatial mapping of the T exhaustion scores. j, UMAP of T cells showing the exhaustion scores. k, UMAP of T cells showing the spatial K-distances to their 15 nearest tumor cells. l, Boxplot comparing the T cell exhaustion scores between different T cell states (CD8T n = 455; NaïveT n = 144; CD4T n = 714; CD4Te n = 97; Treg n = 313; CD8Te n = 158). m, Boxplot comparing the T cell exhaustion scores between T cells proximal to tumor cells (TP) and T cells distal to tumor cells (TD) (TP n = 1,631; TD n = 250). n, Boxplot comparing the T cell exhaustion scores between TP and TD in each T cell state. In l, m and n, two-sided Wilcoxon rank sum tests were performed. Boxplots show the median with interquartile ranges (25–75%); whiskers extend to 1.5× the interquartile range from the box.
Article
Single-cell RNA sequencing methods can profile the transcriptomes of single cells but cannot preserve spatial information. Conversely, spatial transcriptomics assays can profile spatial regions in tissue sections, but do not have single-cell resolution. Here, we developed a computational method called CellTrek that combines these two datasets to achieve single-cell spatial mapping through coembedding and metric learning approaches. We benchmarked CellTrek using simulation and in situ hybridization datasets, which demonstrated its accuracy and robustness. We then applied CellTrek to existing mouse brain and kidney datasets and showed that CellTrek can detect topological patterns of different cell types and cell states. We performed single-cell RNA sequencing and spatial transcriptomics experiments on two ductal carcinoma in situ tissues and applied CellTrek to identify tumor subclones that were restricted to different ducts, and specific T cell states adjacent to the tumor areas. Our data show that CellTrek can accurately map single cells in diverse tissue types to resolve their spatial organization.
 
Article
Living bacteria therapies have been proposed as an alternative approach to treating a broad array of cancers. In this study, we developed a genetically encoded microbial encapsulation system with tunable and dynamic expression of surface capsular polysaccharides that enhances systemic delivery. Based on a small RNA screen of capsular biosynthesis pathways, we constructed inducible synthetic gene circuits that regulate bacterial encapsulation in Escherichia coli Nissle 1917. These bacteria are capable of temporarily evading immune attack, whereas subsequent loss of encapsulation results in effective clearance in vivo. This dynamic delivery strategy enabled a ten-fold increase in maximum tolerated dose of bacteria and improved anti-tumor efficacy in murine models of cancer. Furthermore, in situ encapsulation increased the fraction of microbial translocation among mouse tumors, leading to efficacy in distal tumors. The programmable encapsulation system promises to enhance the therapeutic utility of living engineered bacteria for cancer.
 
M⁶A-SAC-seq strategy and development
a, MjDim1 uses allylic-SAM as a cofactor to label m⁶A to a⁶m⁶A, which undergoes cyclization following I2 treatment. b, An m⁶A-modified 12-mer RNA probe was treated with MjDim1 and allylic-SAM, followed by matrix-assisted laser desorption ionization (MALDI) characterization. The added molecular weight is that of the allyl group. c, An m⁶A-free 12-mer RNA probe was treated with MjDim1 and allylic-SAM, followed by MALDI characterization. No detectable new product appeared. d, Michaelis–Menten steady-state kinetics of the MjDim1-catalyzed allyl transfer to an m⁶A-containing probe (MALDI_Probe_m⁶A in Supplementary Table 1). Data are represented as mean ± s.e.m. for two biological replicates × two technical replicates. e, Michaelis–Menten steady-state kinetics of the MjDim1-catalyzed allyl transfer to an unmodified control probe (MALDI_Probe_A in Supplementary Table 1). Data are represented as mean ± s.e.m. for two biological replicates × two technical replicates. f, Cyclized a⁶m⁶A induces higher mutation rates than cyclized a⁶A in various RNA sequence contexts when using HIV RT. RNA oligonucleotides containing a⁶A or a⁶m⁶A were synthesized by incorporating O⁶-phenyl-adenosine phosphoramidite into the designed sequence containing an NNXNN motif (X = a⁶A or a⁶m⁶A).
Characteristics of quantitative m⁶A maps in poly(A)-tailed RNAs from HeLa, HEK293 and HepG2 cells
a, Number (bar plots) and modification fractions (violin box plots) distribution of m⁶A sites in different RNA regions that include 3′-UTR, CDS, intronic, 5′-UTR, intergenic and promoter. In box plots, lower and upper hinges represent first and third quartiles, the center line represents the median, the red dot represents the mean, and whiskers represent ±1.5× the interquartile range (HEK293: n = 6,201 3′-UTR, n = 4,967 CDS, n = 514 5′-UTR, n = 379 intronic, n = 148 intergenic, n = 25 promoter; HeLa: n = 6,004 3′-UTR, n = 4,069 CDS, n = 377 5′-UTR, n = 324 intronic, n = 103 intergenic, n = 15 promoter; HepG2: n = 6,030 3′-UTR, n = 4,585 CDS, n = 460 5′-UTR, n = 126 intronic, n = 57 intergenic, n = 12 promoter). b, k-means clustering was performed for RNA 100-nt bins that contain m⁶A sites in at least one of the three cell lines. The 5′-UTR, 3′-UTR and CDS 100-nt bins were classified into seven clusters by k-means clustering on the m⁶A fractions sum of 100-nt bin. Each row represents a 100-nt bin, and the number of 100-nt bins is shown in the bracket. c, GO Biological Process (GOBP) enrichment analysis for the m⁶A clusters defined in b. Cell type-specific clusters (C5, C13–C15 and C24–C26) are highlighted by bold italic.
Effects of m⁶A on the modified RNAs in cell lines
a, Cumulative curves and box violin plots showing the distribution of transcript lifetime with different m⁶A fractions in HeLa cells. Transcripts were classified into three groups (high, medium and low) with equal numbers of transcripts based on the sum of their m⁶A fractions (left: n = 1,576 high, n = 1,575 medium, n = 1,575 low, n = 5,277 no m⁶A; right: n = 1,672 high, n = 1,671 medium, n = 1,671 low, n = 6,739 no m⁶A). Two lifetime data sets, GSE98856 (left) and GSE49339 (right), were used to confirm each other. The P value between two groups (the no m⁶A group as reference) was determined by one-tailed Wilcoxon rank-sum test in the violin box plots. b, Cumulative curves and box violin plots showing the lifetime distribution of 5′-UTR-m⁶A-only (n = 77), 3′-UTR-m⁶A-only (n = 1,579), CDS-m⁶A-only (n = 889) and non-m⁶A (n = 5,277) transcripts; lifetime data set: GSE98856. c, Cumulative curves and box violin plots showing the distribution of lifetime for 5′-UTR-m⁶A-only (n = 110), 3′-UTR-m⁶A-only (n = 2,200), CDS-m⁶A-only (n = 1,182) and non-m⁶A (n = 6,739) transcripts in siControl (control short interfering RNA) and siYTHDF2 (YTHDF2 short interfering RNA) data sets and their ratios in HeLa cells. The y axis label of the box violin plot is the same as the x axis label of the cumulative curve plot; lifetime data set: GSE49339. d, Distribution of MFE of RNA folding for a 31-nt sliding window with or without m⁶A on the 16th adenosines. The upstream and downstream 15 nt of adenines with or without m⁶A were used to generate the 31-mer window. All identified m⁶A sites (n = 12,234 HEK293, n = 10,892 HeLa, n = 11,270 HepG2) were used to calculate MFE, and the number of non-m⁶A sites was 10,000 for each cell type. e, Distribution of PhastCons scores of m⁶A and non-m⁶A sites in cell lines. In different cell lines, m⁶A sites were classified into two categories (low and high) based on their PhastCons scores (low ≤ 0.5, high > 0.5) (HEK293: n = 2,711 low, n = 7,074 high; HeLa: n = 2,482 low, n = 6,139 high; HepG2: n = 2,230 low, n = 6,998 high). For b–e, the P value was determined by one-tailed Wilcoxon rank-sum test. In box plots (a–e), lower and upper hinges represent first and third quartiles, the center line represents the median, the red dot represents the mean, and whiskers represent ±1.5× the interquartile range.
M⁶A dynamics across hematopoietic stem cell differentiation into monocytes
a, Distribution of m⁶A fractions in different genomic features. Colors indicate time points. In box plots, lower and upper hinges represent first and third quartiles, the center line represents the median, and whiskers represent ±1.5× the interquartile range (5′-UTR: n = 867 d0, n = 773 d3, n = 628 d6, n = 647 d9; CDS: n = 11,630 d0, n = 12,277 d3, n = 7,215 d6, n = 7,603 d9; 3′-UTR: n = 15,243 d0, n = 15,393 d3, n = 10,378 d6, n = 11,310 d9; intron: n = 3,611 d0, n = 1,416 d3, n = 1,481 d6, n = 3,560 d9; ncRNA: n = 1,473 d0, n = 1,232 d3, n = 948 d6, n = 1,073 d9; intergenic region: n = 590 d0, n = 424 d3, n = 414 d6, n = 581 d9). b, The log2-transformed fold changes (log2 FC) of expression levels of transcripts with dynamic or stable m⁶A sites in different feature regions of mRNAs between adjacent time points are shown. The dynamic or stable m⁶A sites were defined as those that were detected between adjacent time points, and their changes of m⁶A fractions were more or less than 20%, respectively. Colors indicate the dynamic or stable status. P values were determined using a two-tailed Mann–Whitney U-test; NS, not significant. In box plots, lower and upper hinges represent first and third quartiles, the center line represents the median, and whiskers represent ±1.5× the interquartile range (5′-UTR: n = 613, n = 299, n = 576, n = 259, n = 531, n = 254; CDS: n = 3,341, n = 2,220, n = 3,124, n = 1,767, n = 2,814, n = 1,611; 3′-UTR: n = 3,586, n = 2,740, n = 3,380, n = 2,304, n = 3,263, n = 2,171; intron: n = 1,558, n = 253, n = 1,050, n = 198, n = 1,657, n = 290). The order of the n number in each region is consistent with that of box plots shown in each region. c, Alluvial plots showing global m⁶A dynamics on feature regions of mRNAs during monocytopoiesis. Each line represents one transcript bearing m⁶A at different transcript regions across four time points. Colors indicate feature regions where m⁶A was initially installed at d0. d, Number of genes that gained or lost m⁶A in different regions of mRNAs were counted by comparing to the previous time point. Only genes gaining or losing m⁶A on the specific genomic region were considered. Colors indicate the status of gain or loss. e, Heat maps depicting clusters of m⁶A sites by m⁶A stoichiometries in different transcript regions of mRNAs. The dendrogram in each cluster was constructed using complete linkage based on Euclidean distance. Numbers near the dendrograms represent cluster identifiers. f, Heat map illustrating GO analysis on genes with dynamic or stable m⁶A in different transcript regions during differentiation. The dendrograms in both rows and columns were constructed using complete linkage based on Euclidean distance. Between adjacent time points, m⁶A sites with a stoichiometry difference more than 20% were defined as dynamic, and those less than 20% were defined as stable.
M⁶A modification impacting gene expression during monocytopoiesis
a, Heat map showing expression profiles of m⁶A-modified and differentially expressed genes (FC > 1.5 or FC < 0.667, FDR < 0.05) by comparing any two time points during HSPC differentiation. Gene expression levels were scaled by z score across time points. The dendrogram was constructed using complete linkage based on Euclidean distance. Master TFs involved in regulating HSPC differentiation into monocytes are labeled in the heat map; lincRNA, long intergenic non-coding RNA; sncRNA, small non-coding RNA. b, Examples displaying changes of both m⁶A stoichiometries in different transcript regions and expression levels of key TF transcripts; CPM, counts per million. c, Distribution of counts of genes with changes in both m⁶A stoichiometries in different transcript regions and expression levels between adjacent time points. Colors indicate changes, with hyper- or hypomethylation defined as having methylation difference (MD) > 10% or MD < –10%, respectively; up- or downregulated gene expression differences are defined as having FDR < 0.05 and FC > 1.5 or FC < 0.667, respectively. d, Heat map showing TF-regulated gene enrichment analysis on the gene sets with m⁶A in the CDS and 3′-UTR in c. The dendrograms in both rows and columns were constructed using complete linkage based on Euclidean distance. Master TFs during differentiation were labeled in the heat map. e, Heat map (left) depicting the overlap between RBP eCLIP peaks and m⁶A sites and bar plots (right) showing the distribution of counts of their shared binding regions, respectively. The dendrogram in rows was constructed using complete linkage based on Euclidean distance. The percent overlap was represented by the average of the fractions of overlap that was reciprocal for RNA m⁶A sites and RBP eCLIP peaks. The co-occurrence criteria between RBPs and m⁶A was defined as both of RBP and m⁶A reciprocal fractions of site overlap > 5% and one of them > 9% at least at one time point.
Article
Functional studies of the RNA N⁶-methyladenosine (m⁶A) modification have been limited by an inability to map individual m⁶A-modified sites in whole transcriptomes. To enable such studies, here, we introduce m⁶A-selective allyl chemical labeling and sequencing (m⁶A-SAC-seq), a method for quantitative, whole-transcriptome mapping of m⁶A at single-nucleotide resolution. The method requires only ~30 ng of poly(A) or rRNA-depleted RNA. We mapped m⁶A modification stoichiometries in RNA from cell lines and during in vitro monocytopoiesis from human hematopoietic stem and progenitor cells (HSPCs). We identified numerous cell-state-specific m⁶A sites whose methylation status was highly dynamic during cell differentiation. We observed changes of m⁶A stoichiometry as well as expression levels of transcripts encoding or regulated by key transcriptional factors (TFs) critical for HSPC differentiation. m⁶A-SAC-seq is a quantitative method to dissect the dynamics and functional roles of m⁶A sites in diverse biological processes using limited input RNA.
 
Article
Although several monoclonal antibodies (mAbs) targeting severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have been approved for coronavirus disease 2019 (COVID-19) therapy, development was generally inefficient, with lead generation often requiring the production and testing of numerous antibody candidates. Here, we report that the integration of target–ligand blocking with a previously described B cell receptor-sequencing approach (linking B cell receptor to antigen specificity through sequencing (LIBRA-seq)) enables the rapid and efficient identification of multiple neutralizing mAbs that prevent the binding of SARS-CoV-2 spike (S) protein to angiotensin-converting enzyme 2 (ACE2). The combination of target–ligand blocking and high-throughput antibody sequencing promises to increase the throughput of programs aimed at discovering new neutralizing antibodies. B cell receptor sequencing with ligand blocking speeds up neutralizing antibody discovery.
 
Article
Extending the success of cellular immunotherapies against blood cancers to the realm of solid tumors will require improved in vitro models that reveal therapeutic modes of action at the molecular level. Here we describe a system, called BEHAV3D, developed to study the dynamic interactions of immune cells and patient cancer organoids by means of imaging and transcriptomics. We apply BEHAV3D to live-track >150,000 engineered T cells cultured with patient-derived, solid-tumor organoids, identifying a ‘super engager’ behavioral cluster comprising T cells with potent serial killing capacity. Among other T cell concepts we also study cancer metabolome-sensing engineered T cells (TEGs) and detect behavior-specific gene signatures that include a group of 27 genes with no previously described T cell function that are expressed by super engager killer TEGs. We further show that type I interferon can prime resistant organoids for TEG-mediated killing. BEHAV3D is a promising tool for the characterization of behavioral-phenotypic heterogeneity of cellular immunotherapies and may support the optimization of personalized solid-tumor-targeting cell therapies. The dynamics and molecular mechanisms of T cell therapies are probed in cancer organoids.
 
Article
The use of therapeutic monoclonal antibodies is constrained because single antigen targets often do not provide sufficient selectivity to distinguish diseased from healthy tissues. We present HexElect®, an approach to enhance the functional selectivity of therapeutic antibodies by making their activity dependent on clustering after binding to two different antigens expressed on the same target cell. lmmunoglobulin G (lgG)-mediated clustering of membrane receptors naturally occurs on cell surfaces to trigger complement- or cell-mediated effector functions or to initiate intracellular signaling. We engineer the Fc domains of two different lgG antibodies to suppress their individual homo-oligomerization while promoting their pairwise hetero-oligomerization after binding co-expressed antigens. We show that recruitment of complement component C1q to these hetero-oligomers leads to clustering-dependent activation of effector functions such as complement mediated killing of target cells or activation of cell surface receptors. HexElect allows selective antibody activity on target cells expressing unique, potentially unexplored combinations of surface antigens. Antibody pairs are made to act selectively on cells co-expressing two targets by engineering their Fc domains.
 
Article
Mutations in Ras family proteins are implicated in 33% of human cancers, but direct pharmacological inhibition of Ras mutants remains challenging. As an alternative to direct inhibition, we screened for sensitivities in Ras-mutant cells and discovered 249C as a Ras-mutant selective cytotoxic agent with nanomolar potency against a spectrum of Ras-mutant cancers. 249C binds to vacuolar (V)-ATPase with nanomolar affinity and inhibits its activity, preventing lysosomal acidification and inhibiting autophagy and macropinocytosis pathways that several Ras-driven cancers rely on for survival. Unexpectedly, potency of 249C varies with the identity of the Ras driver mutation, with the highest potency for KRASG13D and G12V both in vitro and in vivo, highlighting a mutant-specific dependence on macropinocytosis and lysosomal pH. Indeed, 249C potently inhibits tumor growth without adverse side effects in mouse xenografts of KRAS-driven lung and colon cancers. A comparison of isogenic SW48 xenografts with different KRAS mutations confirmed that KRASG13D/+ (followed by G12V/+) mutations are especially sensitive to 249C treatment. These data establish proof-of-concept for targeting V-ATPase in cancers driven by specific KRAS mutations such as KRASG13D and G12V. Cancers with common KRAS mutations can be treated with an inhibitor of lysosomal acidification in mice.
 
Article
Many biomedical questions demand scalable, deep, and accurate proteome analysis of small samples, including single cells. A scalable framework of multiplexed data-independent acquisition for mass spectrometry enables time saving by parallel analysis of both peptide ions and protein samples, thereby realizing multiplicative gains in throughput.
 
Article
An influenza vaccine is created by attenuating the live virus through targeted proteolysis.
 
Article
The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants with potential resistance to existing drugs emphasizes the need for new therapeutic modalities with broad variant activity. Here we show that ensovibep, a trispecific DARPin (designed ankyrin repeat protein) clinical candidate, can engage the three units of the spike protein trimer of SARS-CoV-2 and inhibit ACE2 binding with high potency, as revealed by cryo-electron microscopy analysis. The cooperative binding together with the complementarity of the three DARPin modules enable ensovibep to inhibit frequent SARS-CoV-2 variants, including Omicron sublineages BA.1 and BA.2. In Roborovski dwarf hamsters infected with SARS-CoV-2, ensovibep reduced fatality similarly to a standard-of-care monoclonal antibody (mAb) cocktail. When used as a single agent in viral passaging experiments in vitro, ensovibep reduced the emergence of escape mutations in a similar fashion to the same mAb cocktail. These results support further clinical evaluation of ensovibep as a broad variant alternative to existing targeted therapies for Coronavirus Disease 2019 (COVID-19).
 
Article
Circular RNAs (circRNAs) are stable and prevalent RNAs in eukaryotic cells that arise from back-splicing. Synthetic circRNAs and some endogenous circRNAs can encode proteins, raising the promise of circRNA as a platform for gene expression. In this study, we developed a systematic approach for rapid assembly and testing of features that affect protein production from synthetic circRNAs. To maximize circRNA translation, we optimized five elements: vector topology, 5′ and 3′ untranslated regions, internal ribosome entry sites and synthetic aptamers recruiting translation initiation machinery. Together, these design principles improve circRNA protein yields by several hundred-fold, provide increased translation over messenger RNA in vitro, provide more durable translation in vivo and are generalizable across multiple transgenes.
 
Variant quantification
a, Location of Austrian WWTPs (grey) and those included in the surveillance programme (black). WWTPs labelled with names are used to showcase analysis details throughout the manuscript. These WWTPs were selected for their size and geographic position, namely, a small and a large WWTP from west, south and east. b, Results of the VaQuERo variant quantification analysis for the ten most abundant variants for all WWTPs with more than 17 time points. c, Spatiotemporal spread of selected variants. Time is projected on the vertical axis; position along the two main transportation axes (west-east and east-south mobility axis) is projected on the horizontal axis. The colour indicates relative variant abundance of B.1.160, B.1.258, Alpha (B.1.1.7), Delta (B.1.617.2) and Omicron (BA.1 and BA.2), from top left to bottom right, as deduced from the measured points and interpolated with B-splines. White crosses indicate measured data points.
Validation
Comparison between relative variant abundance as deduced from WW variant surveillance and from the aggregated statistics of the epidemiological case surveillance. Single data points are aggregated over a week and catchment areas. a, Detailed comparison between relative variant frequencies deduced from WW and from all variant-typed individual cases in the catchment of selected WW treatment plants (WWTP) from west, central and eastern Austria with a small and a large population size, respectively. Black line represents the signal from the WW. The capped bars (lollipop plot) represent the case surveillance records, whereas the respective colour indicates the range of absolute case numbers per variant for the respective catchment area and week. The chosen breaks correspond to the 0, 0.2, 0.5 and 0.75 quantile of the observed zero-truncated absolute case count distribution. b, Enumeration of the agreement and disagreement between case-based and WW-based surveillance with respect to the detection of each variant in a catchment in the same week. Detection is defined by a relative frequency >0. Below the associated Cohen’s κ coefficient, testing a one-sided alternative hypothesis, its 95% confidence interval and corresponding multiple testing corrected P values are depicted. c, Like the inter-rater reliability test for detection presented in b but examining the agreement to identify dominant variant with a relative frequency >0.5. d, Rank correlation across all detected variants and all surveyed WWTPs. To reduce bias caused by the bimodal distribution of relative frequencies with many values close to 0 or 1, values <0.1 and >0.9 were removed before the analysis. e, Absolute case counts and the relative case counts for all data points of the case-based monitoring programme for which no corresponding variant-specific signal was detected in the WW. Black bars depict distribution of both variables, black lines the corresponding medians.
Analysis of mutation patterns
a, Mutation constellations in tree representation as clustered by their mutation frequencies in weekly sample batches from all Carinthian samples in the period from December 2020 to February 2021. The first constellation enriched with Alpha mutations is showcased with the frequencies of the mutations as observed in the samples used for deconvolution (heat map, left) and the frequency of the respective mutation in all samples of a specific variant as deposited in GISAID (heat map, right). b, Comparison of the relative abundance of Alpha, Delta and Omicron variants and the observed nucleotide diversity π in the WW samples. WWTPs with more than three pre-Alpha data points are shown individually (bottom), all WWTP are shown cumulatively (top). To test for reduction of nucleotide diversity during emergence and dominance of the occurrent variants, a one-sided Mann-Whitney U test between variant introduction (dotted vertical line) and end of dominance (dot-dashed vertical line) was performed. Corresponding P values are indicated.
Quantitative trend analysis of virus load interlaced with variant quantification
a, Distribution of WW deduced reproduction number Rww for Alpha (B.1.1.7), Delta (B.1.617.2), Omicron (BA.2) and all other variants as calculated by an integration of absolute case estimate by RT-qPCR and variant estimate by sequencing, showcased for six WWTPs across the time period from January 2021 to February 2022. Depicted distribution are deduced from between 38 to 242 data points. b, Ratios of the variant-specific Rww from the same time point and the same WWTP exhibit a systematic shift to the values >1, with a median of 1.38 between Alpha and pre-Alpha and of 1.85 between Omicron and Delta, indicating a 38% increased transmissibility of Alpha over other competing variants and of 85% of Omicron over Delta, when subjected to an equal environment. Depicted distribution are deduced from between 25 to 96 data points. Horizontal box plots indicate 25th, 50th and 75th percentile (boxes) and up to the 1.5 times interquartile range contiguous from there (whiskers).
Article
SARS-CoV-2 surveillance by wastewater-based epidemiology is poised to provide a complementary approach to sequencing individual cases. However, robust quantification of variants and de novo detection of emerging variants remains challenging for existing strategies. We deep sequenced 3,413 wastewater samples representing 94 municipal catchments, covering >59% of the population of Austria, from December 2020 to February 2022. Our system of variant quantification in sewage pipeline designed for robustness (termed VaQuERo) enabled us to deduce the spatiotemporal abundance of predefined variants from complex wastewater samples. These results were validated against epidemiological records of >311,000 individual cases. Furthermore, we describe elevated viral genetic diversity during the Delta variant period, provide a framework to predict emerging variants and measure the reproductive advantage of variants of concern by calculating variant-specific reproduction numbers from wastewater. Together, this study demonstrates the power of national-scale WBE to support public health and promises particular value for countries without extensive individual monitoring. Wastewater surveillance of SARS-CoV-2 at the national scale tracks emerging variants.
 
Article
Current mass spectrometry methods enable high-throughput proteomics of large sample amounts, but proteomics of low sample amounts remains limited in depth and throughput. To increase the throughput of sensitive proteomics, we developed an experimental and computational framework, called plexDIA, for simultaneously multiplexing the analysis of peptides and samples. Multiplexed analysis with plexDIA increases throughput multiplicatively with the number of labels without reducing proteome coverage or quantitative accuracy. By using three-plex non-isobaric mass tags, plexDIA enables quantification of threefold more protein ratios among nanogram-level samples. Using 1-hour active gradients, plexDIA quantified ~8,000 proteins in each sample of labeled three-plex sets and increased data completeness, reducing missing data more than twofold across samples. Applied to single human cells, plexDIA quantified ~1,000 proteins per cell and achieved 98% data completeness within a plexDIA set while using ~5 minutes of active chromatography per cell. These results establish a general framework for increasing the throughput of sensitive and quantitative protein analysis. Proteomics of small sample sizes using data-independent acquisition methods achieves higher throughput with multiplexing.
 
Article
The seemingly simple modular technology for making new mRNA vaccines could make vaccine developers a victim of their own success.
 
The concept of an RDD-based analysis workflow
a, Perform spectral alignment of the MS/MS-based untargeted metabolomics data from human biospecimens with data from reference samples that have controlled vocabularies for metadata. This can, optionally, be combined with MS/MS libraries. b, Link the spectral matches to the source information from the metadata from the reference samples. Create a data table of source ontology, human biospecimen and counts to enable data science and interpretation.
RDD with food reference data
a, Food RDD analysis schema. (int. = intensity) b, Food spectral counts (1% FDR²¹) observed in plasma from a sleep restriction and circadian misalignment study that controlled the diet of the participants (n = 371 samples from 20 healthy adults)¹⁸. The size of node represents the relative number of spectral matches at each food level. Blue arrows indicate foods that could be explained although they were not provided in the study; orange arrow indicate source is not known. c, A crossover experiment between centenarian data from Italy and a sleep and circadian study from the US, for both fecal and plasma samples. Study-region-specific foods consumed by those individuals (yes) versus a different set of study-region-specific foods (no). One-way Welch’s t-test, thick line is the mean, range within the box is the interquartile range (IQR) from the 25th to 75th quartile, whiskers indicate the minimum and maximum. d, PCA of food counts color coded by vegan (brown) versus omnivore data (green). e, Statistical analysis for the food counts at level 3 of the ontology, in relation to omnivore and vegan data (left six panels, dairy, meat, seafood, legume, fleshy fruit, vegetable, Wilcoxon test, n = 36, 19 are vegan and 19 are omnivore). f, As in e but level 4 ontology using unique spectral counts (spectral usage is the percentage of MS/MS spectra used in the analysis. As they are unnamed ontologies as one would find in microorganism phylogeny in microbiome science (for example kingdom, genus, species) we have denoted these as layers (Right six panels, cow, pig, fish-saltwater, shellfish, citrus, vegetable, Supplementary Table 1). e,f, Boxes represent the IQR; the lower limit is the 25th percentile, the center line is the median, the upper limit is the 75th percentile; bars show the 75th percentile + 1.5 × IQR and the 25th percentile − 1.5 × IQR.
Article
Human untargeted metabolomics studies annotate only ~10% of molecular features. We introduce reference-data-driven analysis to match metabolomics tandem mass spectrometry (MS/MS) data against metadata-annotated source data as a pseudo-MS/MS reference library. Applying this approach to food source data, we show that it increases MS/MS spectral usage 5.1-fold over conventional structural MS/MS library matches and allows empirical assessment of dietary patterns from untargeted data. Metabolomics is improved by using a reference library of both known and unknown molecules.
 
Establishment of the PROTAC virus production system
a, Schematic illustration of the generation of PROTAC viruses that are highly reproductive in TEVp-expressing stable cells but live attenuated by proteasome-mediated viral protein destabilization in conventional cells. VP, viral protein; Ub, ubiquitin. b, Sequences of amino acids and gene of PTD. The proteasome-targeting peptide is depicted in blue, and the TEVcs is depicted in orange. c, Characterization of engineered PROTAC virus-producing stable cell lines by the expression of TEVp measured by qPCR and western blotting. Data are presented as mean ± s.d.; n = 3 biological replicates. d, Comparison of the capabilities of TEVp-expressing stable cell lines and their parental cell lines to propagate WT influenza viruses. Data are presented as mean ± s.d.; n = 3 biological replicates. e, Characterization of the generation of PROTAC viruses by the CPE by inoculating putative viruses in MDCK-TEVp cells. CPE was measured 4–5 days after inoculation by CellTiter-Glo assay. Data are presented as mean ± s.d.; n = 3 biological replicates. f, Characterization of TEVp-dependent replication of M1-PTD by CPE. Cells were infected with M1-PTD or WT virus (MOI = 0.01), and CPE was observed 4 days after infection (n = 3). Scale bar, 200 µm.
Source data
Characterization of PROTAC virus M1-PTD
a, Multicycle replication kinetic curves of M1-PTD in MDCK-TEVp and conventional MDCK.2 cells. Data are presented as mean ± s.d.; n = 3 biological replicates. The detection limit is 70 PFU/mL. b, Comparisons of the plaque phenotypes of M1-PTD and WT virus in MDCK-TEVp (+TEVp) and conventional MDCK.2 (−TEVp) cells. Obvious plaque formation, representing viral propagation in cells, was observed only in MDCK-TEVp cells and not in MDCK.2 cells for M1-PTD (n = 3). c, Immunofluorescence staining of influenza viral M1 at 48 hours after infection (MOI = 0.01) showing robust replication of M1-PTD only in MDCK-TEVp cells (+TEVp) and not in MDCK.2 cells (−TEVp). Representative images are shown for each condition (n = 3). Green, M1; blue, nuclei; scale bar, 100 µm. d, Immunofluorescence staining of influenza viral NP at 48 hours after infection (MOI = 0.01) showing robust replication of M1-PTD only in MDCK-TEVp cells (+TEVp) and not in MDCK.2 cells (−TEVp). Representative images are shown for each condition (n = 3). Green, NP; blue, nuclei; scale bar, 100 µm. e,f, Immunofluorescence staining of influenza viral M1 and NP staining at 48 hours after infection (MOI = 0.01) showing the replication competence of M1-PTD in conventional MDCK.2 cells in the absence and presence of MG-132 (e). Data represent the fold change of infected cells (M1-positive and NP-positive) by quantitatively analyzing three representative areas from each condition (f). Data are presented as mean ± s.d.; n = 3 biological replicates; one-way ANOVA with Dunnett’s multiple comparisons test. Green, M1 or NP; blue, nuclei; scale bar, 100 µm. g, PTD-mediated viral M1 protein degradation is dependent on the proteasome. Conventional MDCK.2 cells were infected with M1-PTD or WT viruses (MOI = 0.1) in the absence and presence of proteasome inhibitor MG-132 (50 nM) and analyzed for viral M1 protein levels by western blotting at indicated times (n = 3). h, PTD-mediated viral M1 protein degradation is dependent on the VHL E3 ubiquitin ligase. Conventional MDCK.2 cells were infected with M1-PTD (MOI = 0.1) in the absence and presence of 200 µM VH298, an inhibitor of VHL E3 ubiquitin ligase, and analyzed for viral M1 protein levels by western blotting at 48 hours after infection (n = 3).
Source data
Evaluation of the in vivo safety of PROTAC virus M1-PTD in mice and ferrets
a, Survival rates and body weights of mice (n = 10) after intranasal infection with the indicated viruses. Data are presented as mean ± s.d. b, Viral titers in mouse tissues (n = 5) at day 3 after infection with 10⁵ PFU of indicated viruses. Data are plotted for individual mice and overlaid with mean ± s.d.; unpaired two-tailed t-test. c, Viral titers in different organs of ferrets (n = 3) at day 3 after infection with 10⁵ PFU of WT WSN or M1-PTD. Data are plotted for individual ferrets and overlaid with mean ± s.d.; unpaired two-tailed t-test.
Source data
Evaluation of the immunogenicity of PROTAC virus M1-PTD in mice and ferrets
a, HI and NT antibody responses in mouse sera at day 21 after vaccination (n = 5). Data are plotted for individual mice and overlaid with means ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. b, Viral surface protein HA or internal protein NP-specific IgG antibody responses in mouse sera at day 21 after vaccination (n = 5). Data are plotted for individual mice and overlaid with means ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. c, Viral NP-specific mucosal IgA antibody responses in mouse lungs at day 21 after vaccination (n = 5). Data are plotted for individual mice and overlaid with means ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. d, Viral NP antigen-specific CD8 T cell responses in the lungs of mice at day 21 after vaccination (n = 5). Data are plotted for individual mice and overlaid with means ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. e, Viral M1 antigen-specific T cell responses in the lungs of mice measured by ELISpot assay (n = 5). Two M1 peptides, M1128–135 MGLIYNRM (left) and M158–66 GILGFVFTL (right), were used as the stimuli. IFNγ-expressing cells per million cells are shown. Data are presented as mean ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. f,g, Antibody responses in vaccinated ferrets. Female ferrets at the age of 4–6 months were inoculated with 10⁶ PFU of the indicated vaccines. Three weeks after vaccination, sera were collected for detection of HI and NT antibody titers (n = 6) (f), and lungs were collected for detection of virus-specific mucosal IgA antibody titers (n = 3) (g).
Source data
Characterizations of the protective efficacy of M1-PTD in mice and ferrets
a, Viral titers at day 3 after challenge with WT WSN virus in the lungs of mice vaccinated with the indicated vaccines (n = 5). Data are plotted for individual mice and overlaid with mean ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. The detection limit is 70 PFU/mL. b, Survival rates and body weights of vaccinated mice (n = 5) after challenge with WT WSN virus. Data are presented as mean ± s.d. c, Protective efficacy of M1-PTD against homologous virus challenge in ferrets. Vaccinated ferrets with indicated vaccines were challenged with 10⁷ PFU of WT WSN viruses. Lungs were collected 3 days after challenge for detection of viral titers (n = 3). Data are plotted for individual ferrets and overlaid with mean ± s.d.; one-way ANOVA with Tukey’s multiple comparisons test. d, Cross-reactive protection efficacy of M1-PTD against heterologous virus challenge in ferrets. Viral titers in the nasal washes and lungs of vaccinated ferrets with 10⁶ PFU of M1-PTD at day 3 after challenge with 10⁶ PFU of heterologous influenza A/Netherlands/602/2009 (pdmH1N1) virus (n = 3). Data are plotted for individual ferrets and overlaid with mean ± s.d.; unpaired two-tailed t-test.
Source data
Article
The usefulness of live attenuated virus vaccines has been limited by suboptimal immunogenicity, safety concerns or cumbersome manufacturing processes and techniques. Here we describe the generation of a live attenuated influenza A virus vaccine using proteolysis-targeting chimeric (PROTAC) technology to degrade viral proteins via the endogenous ubiquitin–proteasome system of host cells. We engineered the genome of influenza A viruses in stable cell lines engineered for virus production to introduce a conditionally removable proteasome-targeting domain, generating fully infective PROTAC viruses that were live attenuated by the host protein degradation machinery upon infection. In mouse and ferret models, PROTAC viruses were highly attenuated and able to elicit robust and broad humoral, mucosal and cellular immunity against homologous and heterologous virus challenges. PROTAC-mediated attenuation of viruses may be broadly applicable for generating live attenuated vaccines. Influenza virus is attenuated for vaccine production using PROTAC degradation technology.
 
Article
Systematically identifying synergistic combinations of targeted agents and immunotherapies for cancer treatments remains difficult. In this study, we integrated high-throughput and high-content techniques—an implantable microdevice to administer multiple drugs into different sites in tumors at nanodoses and multiplexed imaging of tumor microenvironmental states—to investigate the tumor cell and immunological response signatures to different treatment regimens. Using a mouse model of breast cancer, we identified effective combinations from among numerous agents within days. In vivo studies in three immunocompetent mammary carcinoma models demonstrated that the predicted combinations synergistically increased therapeutic efficacy. We identified at least five promising treatment strategies, of which the panobinostat, venetoclax and anti-CD40 triple therapy was the most effective in inducing complete tumor remission across models. Successful drug combinations increased spatial association of cancer stem cells with dendritic cells during immunogenic cell death, suggesting this as an important mechanism of action in long-term breast cancer control. Testing of combinations of cancer immunotherapies and conventional drugs yields promising leads.
 
Recombination-based gene circuits built with plant-compatible parts can control plant gene expression
a, Schematic of recombination-based identify gene circuit construct design (YES gate), before and after recombination has occurred. b, Constructs to identify a repressor of gene expression. The Act2 promoter followed by the TMV 5′ UTR drives the output gene (Rluc), and the TCTP1 promoter drives Fluc, which is used to normalize for transfection efficiency. Test constructs each had either a terminator of transcription or a uORF flanked by FRT recombination sites, upstream of the Rluc coding sequence within the TMV 5′ UTR. c, Test of a range of terminators or uORF, flanked by FRT recombination sites, upstream of the Rluc reporter gene. Circuit output activity is the Rluc/Fluc luminescence ratio (also for e) 24 hours after construct transfection into Arabidopsis protoplasts (n = 4). Crossbar indicates mean; blue bar is the control sample; gray bars represent samples expected to be repressed; and bars with different letters have a significant difference calculated by one-way ANOVA (P = 3.53 × 10⁻¹⁴) and Tukey’s HSD test (P < 0.05). d, Construct diagram of plasmids used to test the 1-input switch (YES gate) design with YES gate symbol below. The region labeled as Test Promoter was individually replaced and tested with NOS, TCTP1, Act2 or 35S promoters. e, Testing of different promoters driving Flp recombinase expression in a YES gate, 24 hours after transfection into Arabidopsis protoplasts (n = 4). Crossbar indicates mean. Bar colors as per c, with green bars representing samples expected to be activated. Bars with different letters have a significant difference calculated by one-way ANOVA (P = 2.35 × 10⁻⁹) and Tukey’s HSD test (P < 0.05). f, Induction of nuclear-localized GFP in Arabidopsis roots after heat shock in transgenic Arabidopsis plants harboring the HSP18.2 promoter driving heat-dependent Flp expression to remove the OCS terminator from the Act2 promoter that blocks GFP transcription. The image is merged from transmitted light and GFP channels and is the maximum intensity projection from ten slices. Scale bar, 100 μm. This experiment was repeated independently three times with similar results each time.
Identification of additional functional recombinases for the construction of complex gene circuits
a, Schematics of the Cre/Flp-based OR gate construct designs. b, Cre/Flp-based OR gate that produces high circuit output activity only when Flp is present alone, 24 hours after transfection into Arabidopsis protoplasts (n = 4), where circuit output activity is the Rluc/Fluc luminescence ratio (as also for d and f). Crossbar indicates mean; blue bar represents the control sample; green bars represent samples expected to be activated (turned on); gray bars represent samples expected to be repressed (turned off), as also for d and f. Bars with different letters have a significant difference calculated by one-way ANOVA (P = 3.49 × 10⁻⁹) and Tukey’s HSD test (P < 0.05). c, Constructs used to test whether the B3 recombinase can adequately remove the OCS terminator and activate Rluc expression, compared to Flp YES gate. d, Test of B3 for YES gate after transfection into Arabidopsis protoplasts, compared to the previously tested Flp-based design. Crossbar indicates mean (n = 4). Bars with different letters have a significant difference calculated by one-way ANOVA (P = 1.94 × 10⁻⁵) and Tukey’s HSD test (P < 0.05). e, Constructs used to test for cross-reactivity between Flp and B3 recombinases. f, Test of cross-reactivity between Flp and B3 recombinases after construct transfection into Arabidopsis protoplasts. Crossbar indicates mean (n = 4). Bars with different letters have a significant difference calculated by one-way ANOVA (P = 3.37 × 10⁻⁸) and Tukey’s HSD test (P < 0.05). g, Summary of cognate and non-cognate recombination activity of Flp and B3. The solid line represents confirmed recombination event, and the dashed line represents a tested, but not active, recombination event.
AND and OR gate construction using recombination-based gene circuits in plants
a, Schematics of the OR gate construct designs. b, A 2-input OR gate that produces high circuit output activity when one of the two inputs is present, 24 hours or 48 hours after Arabidopsis protoplast transfection. Circuit output activity is the Rluc/Fluc luminescence ratio, as also for d (n = 4). Crossbar indicates mean. Bar colors as per Fig. 2. Bars with different letters have a significant difference calculated by one-way ANOVA (24 hours: P = 1.32 × 10⁻⁶; 48 hours: P = 2.12 × 10⁻⁷) and Tukey’s HSD test (P < 0.05). c, Schematics of the AND gate construct designs. d, A 2-input AND gate that produces high Rluc luminescence relative to Fluc luminescence when both of the inputs are present, 24 hours or 48 hours after Arabidopsis protoplast transfection. Circuit output is the Rluc luminescence over the Fluc luminescence (n = 4). Crossbar indicates mean. Bar colors as per Fig. 2. Bars with different letters have a significant difference calculated by one-way ANOVA (24 hours: P = 2.63 × 10⁻⁹; 48 hours: P = 2.1 × 10⁻⁸) and Tukey’s HSD test (P < 0.05). e, AND gate functionality in vivo, driven by condition-specific promoters. Schematic of the construct at top. Confocal images from roots of induced (DEX-treated) and uninduced (DMSO solvent control) transgenic plants. Expression of the output gene (nuclear-localized GFP) is blocked unless in a cell with both recombinases expressed, where Flp expression is induced when exposed to DEX, and B3 expression is driven by the cortex-specific CO2 promoter. Bright-field and GFP (cyan) merged images demonstrate the specificity of activation in response to the application of DEX. To confirm that expression of GFP was in cortex cells, optical sections were created from confocal images of plants treated with cell wall stain PI. The magenta (mCherry and PI) represents the in-focus nuclei (from 35S-driven nuclear-localized mCherry) and cell walls, whereas the cyan (GFP) represents the nuclear-localized induced circuit output in cortex cells. PI signal was used to trace the cell walls (indicated by superimposed colored dashed lines) and demarcate the three outermost root cell type layers (outer to inner): epidermis (ep), cortex (c) and endodermis (end). Scale bars are 100 µm for bright-field and GFP images and 20 µm for optical cross-sections. The experiments with bright-field were repeated four times, each with similar results. The experiments with the PI staining were repeated four times, each with similar results.
Negation (NOT) function implementation in plant cells by output gene CDS or promoter excision
a, Comparison of output (Rluc) repression over time with four different negation (NOT) function designs after transfections into Arabidopsis protoplasts. Left to right: Flp targeting the output gene CDS, B3 targeting the output gene CDS, Flp targeting the output gene promoter and B3 targeting the output gene promoter. Three separate protoplast transfections were measured (24, 48 and 64 hours). Normalized circuit output is the Rluc/Fluc luminescence ratio (circuit output) of the NOT function plasmid, divided by the circuit output of the constitutive Rluc control construct (Fig. 1b), to account differences between plates and timepoints (n = 4, except for constitutive Rluc 48 hours, where n = 3). Crossbar indicates mean. Asterisks denote significance from the post hoc Tukey’s HSD test, performed after one-way ANOVA (FRT generation 1: P = 9.12 × 10⁻⁶; B3RT generation 1: P = 0.000165; FRT generation 2: P = 4.14 × 10⁻⁷; B3RT generation 2: P = 1.75 × 10⁻¹⁰) between the 0-input state and the 1-input state: NS = P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001. Control refers to the activity of the constitutive Rluc construct. b, Circuit output of NOT function plasmids of the 1-input states (recombinase gene present) relative to the 0-input states (no recombinase gene present) over time after transfection into Arabidopsis protoplasts. Generation 1 and Generation 2 refer to NOT function designs where the recombinase targets the output gene CDS or the output gene promoter, respectively. c, A NIMPLY B gate activity 64 hours after transfection of depicted constructs into Arabidopsis protoplasts. Circuit output activity is the Rluc/Fluc luminescence ratio (n = 4). Crossbar indicates mean. Bar colors as per Fig. 2. Bars with different letters have a significant difference calculated by one-way ANOVA (P = 8.96 × 10⁻¹⁴) and Tukey’s HSD test (P < 0.05). NS, not significant.
Construction of AND and NAND gates using split-recombinases
a, Flp split into two recombinase fragments (F1 and F2), each fused to the phage C1 dimerization domain, to construct an AND gate by removal of an FRT-flanked OCS terminator located in the 5′ UTR of the Rluc output gene, measured 24 hours or 48 hours after protoplast transfection. Schematic represents the 2-input construct design, with the circuit state permutations produced by plasmids lacking different combinations of input recombinase fragment(s). Crossbar indicates mean (n = 4). Bar colors as per Fig. 2. Bars with different letters have a significant difference calculated by one-way ANOVA (24 hours: P = 4.03 × 10⁻¹⁴; 48 hours: P = 1.59 × 10⁻¹⁴) and Tukey’s HSD test (P < 0.05). b, Flp split into two recombinase fragments (F1 and F2), each fused to the phage C1 dimerization domain, to construct a NAND gate through removal of the FRT-flanked Act2 promoter upstream of the Rluc output gene, measured 24 hours or 48 hours after protoplast transfection. Schematic represents the 2-input construct design, with the circuit state permutations produced by plasmids lacking different combinations of input recombinase fragment(s). Crossbar indicates mean (n = 4). Bar colors as per Fig. 2. Bars with different letters have a significant difference calculated by one-way ANOVA (24 hours: P = 2.41 × 10⁻⁷; 48 hours: P = 3.45 × 10⁻⁶) and Tukey’s HSD test (P < 0.05).
Article
Plant biotechnology predominantly relies on a restricted set of genetic parts with limited capability to customize spatiotemporal and conditional expression patterns. Synthetic gene circuits have the potential to integrate multiple customizable input signals through a processing unit constructed from biological parts to produce a predictable and programmable output. Here we present a suite of functional recombinase-based gene circuits for use in plants. We first established a range of key gene circuit components compatible with plant cell functionality. We then used these to develop a range of operational logic gates using the identify function (activation) and negation function (repression) in Arabidopsis protoplasts and in vivo, demonstrating their utility for programmable manipulation of transcriptional activity in a complex multicellular organism. Specifically, using recombinases and plant control elements, we activated transgenes in YES, OR and AND gates and repressed them in NOT, NOR and NAND gates; we also implemented the A NIMPLY B gate that combines activation and repression. Through use of genetic recombination, these circuits create stable long-term changes in expression and recording of past stimuli. This highly compact programmable gene circuit platform provides new capabilities for engineering sophisticated transcriptional programs and previously unrealized traits into plants. Transcriptional activity in plants is controlled with a programmable gene circuit.
 
Article
Logic-gated monoclonal antibodies have been designed to improve safety and efficacy.
 
Article
Companies pursue souped-up IL-2 candidates, undeterred by Nektar’s setback.
 
Article
The digitalization of R&D can potentially transform the life sciences, but it cannot succeed without experts collaborating to develop a vision of what a better working environment can look like.
 
Article
Companies are deploying siRNA and antisense oligonucleotides to tackle dangerously high cholesterol driven by genetics, betting that a wider population will benefit.
 
Top-cited authors
Cole Trapnell
  • Harvard University
Gad Getz
  • Broad Institute of MIT and Harvard
Chad Nusbaum
  • Broad Institute of MIT and Harvard
Xian Adiconis
  • Broad Institute of MIT and Harvard
Lin Fan
  • Tongji University