Rebecca J. Carlson
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Rebecca J. Carlson has a total of 48 co-authors. Here are the top co-authors by number of shared publications:
- Nir Hacohen
Massachusetts General Hospital (MGH) Department of Surgery, MGH Krantz Center for Cancer Research, Boston, USA
5 shared publications - Paul C. Blainey
Massachusetts Institute of Technology
5 shared publications - Bingxu Liu
Department of Biochemistry, University of Washington, Seattle, WA 98195, USA
3 shared publications - Matteo Gentili
Broad Institute of MIT and Harvard, Cambridge, MA, USA
3 shared publications Avtar Singh
Broad Institute of MIT and Harvard
2 shared publications- Brian Y. Soong
Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, USA
2 shared publications
Latest Publications
- Open Access
The accessible human surfaceome
Proteins associated with the extracellular face of the plasma membrane ("surface proteins" or the "surfaceome") are important therapeutic targets because they can direct therapeutics to specific cell types and deliver them into cells. However, experimental and computational databases disagree substantially on which proteins make up the human plasma membrane proteome, and none focus specifically on the set of accessible surface proteins that would be useful for therapeutic applications. Here, we used an agentic LLM pipeline to curate the accessible human surfaceome from literature. To assess performance, we built SurfaceBench, a manually curated benchmark of 147 proteins on which five reference databases frequently disagree. Our LLM triage agent classified surface status more accurately than existing databases and recovered 1,108 candidate surface proteins not present in any of them. On held-out clinical and viral-entry surface targets, the pipeline recovered 98–100% of established surface antigens. A literature-anchored deep-dive pipeline run on over 5,000 candidates then returned 1,783 canonical surface proteins and 1,252 with weaker or context-dependent surface evidence. Each queryable record contains 24 LLM-based annotation fields—including confidence, primary localization, and expression breadth—and seven deterministic structural and topological features. Unlike existing databases, our records trace every claim to the agent's reasoning and to exact quotes from the full-text literature it could access. All records can be browsed in an open-source viewer. These results show that LLMs reasoning over unstructured scientific text can build updatable, bespoke databases that outperform expert-curated resources, although performance remains limited by agent access to full-text literature.
Open AccessThe accessible human surfaceome
Proteins associated with the extracellular face of the plasma membrane ("surface proteins" or the "surfaceome") are important therapeutic targets because they can direct therapeutics to specific cell types and deliver them into cells. However, experimental and computational databases disagree substantially on which proteins make up the human plasma membrane proteome, and none focus specifically on the set of accessible surface proteins that would be useful for therapeutic applications. Here, we used an agentic LLM pipeline to curate the accessible human surfaceome from literature. To assess performance, we built SurfaceBench, a manually curated benchmark of 147 proteins on which five reference databases frequently disagree. Our LLM triage agent classified surface status more accurately than existing databases and recovered 1,108 candidate surface proteins not present in any of them. On held-out clinical and viral-entry surface targets, the pipeline recovered 98–100% of established surface antigens. A literature-anchored deep-dive pipeline run on over 5,000 candidates then returned 1,783 canonical surface proteins and 1,252 with weaker or context-dependent surface evidence. Each queryable record contains 24 LLM-based annotation fields—including confidence, primary localization, and expression breadth—and seven deterministic structural and topological features. Unlike existing databases, our records trace every claim to the agent's reasoning and to exact quotes from the full-text literature it could access. All records can be browsed in an open-source viewer. These results show that LLMs reasoning over unstructured scientific text can build updatable, bespoke databases that outperform expert-curated resources, although performance remains limited by agent access to full-text literature.
- Open Access
Classification and functional characterization of regulators of intracellular STING trafficking identified by genome-wide optical pooled screening
STING is an innate immune sensor that traffics across many cellular compartments to carry out its function of detecting cyclic di-nucleotides and triggering defense processes. Mutations in factors that regulate this process are often linked to STING-dependent human inflammatory disorders. To systematically identify factors involved in STING trafficking, we performed a genome-wide optical pooled screen and examined the impact of genetic perturbations on intracellular STING localization. Based on subcellular imaging of STING protein and trafficking markers in 45 million cells perturbed with sgRNAs, we defined 464 clusters of gene perturbations with similar cellular phenotypes. A higher-dimensional focused optical pooled screen on 262 perturbed genes which assayed 11 imaging channels identified 73 finer phenotypic clusters. In a cluster containing USE1, a protein that mediates Golgi to ER transport, we found a gene of unknown function, C19orf25. Consistent with the known role of USE1, loss of C19orf25 enhanced STING signaling. Other clusters contained subunits of the HOPS, GARP and RIC1-RGP1 complexes. We show that HOPS deficiency delayed STING degradation and consequently increased signaling. Similarly, GARP/RIC1-RGP1 loss increased STING signaling by delaying STING exit from the Golgi. Our findings demonstrate that genome-wide genotype-phenotype maps based on high-content cell imaging outperform other screening approaches, and provide a community resource for mining for factors that impact STING trafficking as well as other cellular processes observable in our dataset.
Open AccessClassification and functional characterization of regulators of intracellular STING trafficking identified by genome-wide optical pooled screening
STING is an innate immune sensor that traffics across many cellular compartments to carry out its function of detecting cyclic di-nucleotides and triggering defense processes. Mutations in factors that regulate this process are often linked to STING-dependent human inflammatory disorders. To systematically identify factors involved in STING trafficking, we performed a genome-wide optical pooled screen and examined the impact of genetic perturbations on intracellular STING localization. Based on subcellular imaging of STING protein and trafficking markers in 45 million cells perturbed with sgRNAs, we defined 464 clusters of gene perturbations with similar cellular phenotypes. A higher-dimensional focused optical pooled screen on 262 perturbed genes which assayed 11 imaging channels identified 73 finer phenotypic clusters. In a cluster containing USE1, a protein that mediates Golgi to ER transport, we found a gene of unknown function, C19orf25. Consistent with the known role of USE1, loss of C19orf25 enhanced STING signaling. Other clusters contained subunits of the HOPS, GARP and RIC1-RGP1 complexes. We show that HOPS deficiency delayed STING degradation and consequently increased signaling. Similarly, GARP/RIC1-RGP1 loss increased STING signaling by delaying STING exit from the Golgi. Our findings demonstrate that genome-wide genotype-phenotype maps based on high-content cell imaging outperform other screening approaches, and provide a community resource for mining for factors that impact STING trafficking as well as other cellular processes observable in our dataset.
- Open Access
Single-cell image-based genetic screens systematically identify regulators of Ebola virus subcellular infection dynamics
Ebola virus (EBOV) is a high-consequence filovirus that gives rise to frequent epidemics with high case fatality rates and few therapeutic options. Here, we applied image-based screening of a genome-wide CRISPR library to systematically identify host cell regulators of Ebola virus infection in 39,085,093 million single cells. Measuring viral RNA and protein levels together with their localization in cells identified over 998 related host factors and provided detailed information about the role of each gene across the virus replication cycle. We trained a deep learning model on single-cell images to associate each host factor with predicted replication steps, and confirmed the predicted relationship for select host factors. Among the findings, we showed that the mitochondrial complex III subunit UQCRB is a post-entry regulator of Ebola virus RNA replication, and demonstrated that UQCRB inhibition with a small molecule reduced overall Ebola virus infection with an IC50 of 5 μM. Using a random forest model, we also identified perturbations that reduced infection by disrupting the equilibrium between viral RNA and protein. One such protein, STRAP, is a spliceosome-associated factor that was found to be closely associated with VP35, a viral protein required for RNA processing. Loss of STRAP expression resulted in a reduction in full-length viral genome production and subsequent production of non-infectious virus particles. Overall, the data produced in this genome-wide high-content single-cell screen and secondary screens in additional cell lines and related filoviruses (MARV and SUDV) revealed new insights about the role of host factors in virus replication and potential new targets for therapeutic intervention.
Open AccessSingle-cell image-based genetic screens systematically identify regulators of Ebola virus subcellular infection dynamics
Ebola virus (EBOV) is a high-consequence filovirus that gives rise to frequent epidemics with high case fatality rates and few therapeutic options. Here, we applied image-based screening of a genome-wide CRISPR library to systematically identify host cell regulators of Ebola virus infection in 39,085,093 million single cells. Measuring viral RNA and protein levels together with their localization in cells identified over 998 related host factors and provided detailed information about the role of each gene across the virus replication cycle. We trained a deep learning model on single-cell images to associate each host factor with predicted replication steps, and confirmed the predicted relationship for select host factors. Among the findings, we showed that the mitochondrial complex III subunit UQCRB is a post-entry regulator of Ebola virus RNA replication, and demonstrated that UQCRB inhibition with a small molecule reduced overall Ebola virus infection with an IC50 of 5 μM. Using a random forest model, we also identified perturbations that reduced infection by disrupting the equilibrium between viral RNA and protein. One such protein, STRAP, is a spliceosome-associated factor that was found to be closely associated with VP35, a viral protein required for RNA processing. Loss of STRAP expression resulted in a reduction in full-length viral genome production and subsequent production of non-infectious virus particles. Overall, the data produced in this genome-wide high-content single-cell screen and secondary screens in additional cell lines and related filoviruses (MARV and SUDV) revealed new insights about the role of host factors in virus replication and potential new targets for therapeutic intervention.
- Top Cited PapersOpen Access
- 4 August 2023
- Published by American Association for the Advancement of Science (AAAS) in Science
- Vol. 381 (6657) , 508-514
Human STING is a proton channel
Proton leakage from organelles is a common signal for noncanonical light chain 3B (LC3B) lipidation and inflammasome activation, processes induced upon stimulator of interferon genes (STING) activation. On the basis of structural analysis, we hypothesized that human STING is a proton channel. Indeed, we found that STING activation induced a pH increase in the Golgi and that STING reconstituted in liposomes enabled transmembrane proton transport. Compound 53 (C53), a STING agonist that binds the putative channel interface, blocked STING-induced proton flux in the Golgi and in liposomes. STING-induced LC3B lipidation and inflammasome activation were also inhibited by C53, suggesting that STING’s channel activity is critical for these two processes. Thus, STING’s interferon-induction function can be decoupled from its roles in LC3B lipidation and inflammasome activation.
- 4 August 2023
- Published by American Association for the Advancement of Science (AAAS) in Science
- Vol. 381 (6657) , 508-514
Top Cited PapersOpen AccessHuman STING is a proton channel
Proton leakage from organelles is a common signal for noncanonical light chain 3B (LC3B) lipidation and inflammasome activation, processes induced upon stimulator of interferon genes (STING) activation. On the basis of structural analysis, we hypothesized that human STING is a proton channel. Indeed, we found that STING activation induced a pH increase in the Golgi and that STING reconstituted in liposomes enabled transmembrane proton transport. Compound 53 (C53), a STING agonist that binds the putative channel interface, blocked STING-induced proton flux in the Golgi and in liposomes. STING-induced LC3B lipidation and inflammasome activation were also inhibited by C53, suggesting that STING’s channel activity is critical for these two processes. Thus, STING’s interferon-induction function can be decoupled from its roles in LC3B lipidation and inflammasome activation.
- Open Access
- 12 April 2023
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 120 (16)
A genome-wide optical pooled screen reveals regulators of cellular antiviral responses
The infection of mammalian cells by viruses and innate immune responses to infection are spatiotemporally organized processes. Cytosolic RNA sensors trigger nuclear translocation of the transcription factor interferon regulatory factor 3 (IRF3) and consequent induction of host immune responses to RNA viruses. Previous genetic screens for factors involved in viral sensing did not resolve changes in the subcellular localization of host or viral proteins. Here, we increased the throughput of our optical pooled screening technology by over fourfold. This allowed us to carry out a genome-wide CRISPR knockout screen using high-resolution multiparameter imaging of cellular responses to Sendai virus infection coupled with in situ cDNA sequencing by synthesis (SBS) to identify 80,408 single guide RNAs (sgRNAs) in 10,366,390 cells—over an order of magnitude more genomic perturbations than demonstrated previously using an in situ SBS readout. By ranking perturbations using human-designed and deep learning image feature scores, we identified regulators of IRF3 translocation, Sendai virus localization, and peroxisomal biogenesis. Among the hits, we found that ATP13A1, an ER-localized P5A-type ATPase, is essential for viral sensing and is required for targeting of mitochondrial antiviral signaling protein (MAVS) to mitochondrial membranes where MAVS must be localized for effective signaling through retinoic acid-inducible gene I (RIG-I). The ability to carry out genome-wide pooled screens with complex high-resolution image-based phenotyping dramatically expands the scope of functional genomics approaches.
- 12 April 2023
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 120 (16)
Open AccessA genome-wide optical pooled screen reveals regulators of cellular antiviral responses
The infection of mammalian cells by viruses and innate immune responses to infection are spatiotemporally organized processes. Cytosolic RNA sensors trigger nuclear translocation of the transcription factor interferon regulatory factor 3 (IRF3) and consequent induction of host immune responses to RNA viruses. Previous genetic screens for factors involved in viral sensing did not resolve changes in the subcellular localization of host or viral proteins. Here, we increased the throughput of our optical pooled screening technology by over fourfold. This allowed us to carry out a genome-wide CRISPR knockout screen using high-resolution multiparameter imaging of cellular responses to Sendai virus infection coupled with in situ cDNA sequencing by synthesis (SBS) to identify 80,408 single guide RNAs (sgRNAs) in 10,366,390 cells—over an order of magnitude more genomic perturbations than demonstrated previously using an in situ SBS readout. By ranking perturbations using human-designed and deep learning image feature scores, we identified regulators of IRF3 translocation, Sendai virus localization, and peroxisomal biogenesis. Among the hits, we found that ATP13A1, an ER-localized P5A-type ATPase, is essential for viral sensing and is required for targeting of mitochondrial antiviral signaling protein (MAVS) to mitochondrial membranes where MAVS must be localized for effective signaling through retinoic acid-inducible gene I (RIG-I). The ability to carry out genome-wide pooled screens with complex high-resolution image-based phenotyping dramatically expands the scope of functional genomics approaches.
- Open Access
ESCRT-dependent STING degradation curtails steady-state and cGAMP-induced signaling
STING is an intracellular sensor of cyclic di-nucleotides involved in response to pathogen- or self-derived DNA that induces protective immunity, or if dysregulated, autoimmunity. STING trafficking is tightly linked to its activity. We aimed to systematically characterize genes regulating STING trafficking and to define their impact on STING responses. Based on proximity-ligation proteomics and genetic screens, an ESCRT complex containing HGS, VPS37A and UBAP1 was found to be required for STING degradation and signaling shutdown. Analogous to phosphorylated STING creating a platform for IRF3 recruitment, oligomerization-driven STING ubiquitination by UBE2N formed a platform for ESCRT recruitment at the endosome, responsible for STING signaling shutdown. A UBAP1 mutant that underlies human spastic paraplegia and disrupts ESCRT function led to STING-dependent type I IFN responses at the steady-state, defining ESCRT as a homeostatic regulator of STING signaling.
Open AccessESCRT-dependent STING degradation curtails steady-state and cGAMP-induced signaling
STING is an intracellular sensor of cyclic di-nucleotides involved in response to pathogen- or self-derived DNA that induces protective immunity, or if dysregulated, autoimmunity. STING trafficking is tightly linked to its activity. We aimed to systematically characterize genes regulating STING trafficking and to define their impact on STING responses. Based on proximity-ligation proteomics and genetic screens, an ESCRT complex containing HGS, VPS37A and UBAP1 was found to be required for STING degradation and signaling shutdown. Analogous to phosphorylated STING creating a platform for IRF3 recruitment, oligomerization-driven STING ubiquitination by UBE2N formed a platform for ESCRT recruitment at the endosome, responsible for STING signaling shutdown. A UBAP1 mutant that underlies human spastic paraplegia and disrupts ESCRT function led to STING-dependent type I IFN responses at the steady-state, defining ESCRT as a homeostatic regulator of STING signaling.
- 12 January 2022
- Published by Springer Nature in Nature Protocols
- Vol. 17 (2) , 476-512
Pooled genetic perturbation screens with image-based phenotypes
Discovery of the genetic components underpinning fundamental and disease-related processes is being rapidly accelerated by combining efficient, programmable genetic engineering with phenotypic readouts of high spatial, temporal and/or molecular resolution. Microscopy is a fundamental tool for studying cell biology, but its lack of high-throughput sequence readouts hinders integration in large-scale genetic screens. Optical pooled screens using in situ sequencing provide massively scalable integration of barcoded lentiviral libraries (e.g., CRISPR perturbation libraries) with high-content imaging assays, including dynamic processes in live cells. The protocol uses standard lentiviral vectors and molecular biology, providing single-cell resolution of phenotype and engineered genotype, scalability to millions of cells and accurate sequence reads sufficient to distinguish >106 perturbations. In situ amplification takes ~2 d, while sequencing can be performed in ~1.5 h per cycle. The image analysis pipeline provided enables fully parallel automated sequencing analysis using a cloud or cluster computing environment.
- 12 January 2022
- Published by Springer Nature in Nature Protocols
- Vol. 17 (2) , 476-512
Pooled genetic perturbation screens with image-based phenotypes
Discovery of the genetic components underpinning fundamental and disease-related processes is being rapidly accelerated by combining efficient, programmable genetic engineering with phenotypic readouts of high spatial, temporal and/or molecular resolution. Microscopy is a fundamental tool for studying cell biology, but its lack of high-throughput sequence readouts hinders integration in large-scale genetic screens. Optical pooled screens using in situ sequencing provide massively scalable integration of barcoded lentiviral libraries (e.g., CRISPR perturbation libraries) with high-content imaging assays, including dynamic processes in live cells. The protocol uses standard lentiviral vectors and molecular biology, providing single-cell resolution of phenotype and engineered genotype, scalability to millions of cells and accurate sequence reads sufficient to distinguish >106 perturbations. In situ amplification takes ~2 d, while sequencing can be performed in ~1.5 h per cycle. The image analysis pipeline provided enables fully parallel automated sequencing analysis using a cloud or cluster computing environment.
- Open Access
- 1 October 2018
- Published by Elsevier in Cell Reports
- Vol. 25 (1) , 95-106.e6
IFIT1 Exerts Opposing Regulatory Effects on the Inflammatory and Interferon Gene Programs in LPS-Activated Human Macrophages
- 1 October 2018
- Published by Elsevier in Cell Reports
- Vol. 25 (1) , 95-106.e6
Open AccessIFIT1 Exerts Opposing Regulatory Effects on the Inflammatory and Interferon Gene Programs in LPS-Activated Human Macrophages
- 2 May 2017
- Published by Springer Nature in Applied Microbiology and Biotechnology
- Vol. 101 (14) , 5645-5652
Use of Brevibacillus choshinensis for the production of biologically active brain-derived neurotrophic factor (BDNF)
Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family critical for neuronal cell survival and differentiation, with therapeutic potential for the treatment of neurological disorders and spinal cord injuries. The production of recombinant, bioactive BDNF is not practical in most traditional microbial expression systems because of the inability of the host to correctly form the characteristic cystine-knot fold of BDNF. Here, we investigated Brevibacillus choshinensis as a suitable expression host for bioactive BDNF expression, evaluating the effects of medium type (2SY and TM), temperature (25 and 30 °C), and culture time (48–120 h). Maximal BDNF bioactivity (per unit mass) was observed in cultures grown in 2SY medium at extended times (96 h at 30 °C or >72 h at 25 °C), with resulting bioactivity comparable to that of a commercially available BDNF. For cultures grown in 2SY medium at 25 °C for 72 h, the condition that led to the greatest quantity of biologically active protein in the shortest culture time, we recovered 264 μg/L of BDNF. As with other microbial expression systems, BDNF aggregates did form in all culture conditions, indicating that while we were able to recover biologically active BDNF, further optimization of the expression system could yield still greater quantities of bioactive protein. This study provides confirmation that B. choshinensis is capable of producing biologically active BDNF and that further optimization of culture conditions could prove valuable in increasing BDNF yields.
- 2 May 2017
- Published by Springer Nature in Applied Microbiology and Biotechnology
- Vol. 101 (14) , 5645-5652
Use of Brevibacillus choshinensis for the production of biologically active brain-derived neurotrophic factor (BDNF)
Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family critical for neuronal cell survival and differentiation, with therapeutic potential for the treatment of neurological disorders and spinal cord injuries. The production of recombinant, bioactive BDNF is not practical in most traditional microbial expression systems because of the inability of the host to correctly form the characteristic cystine-knot fold of BDNF. Here, we investigated Brevibacillus choshinensis as a suitable expression host for bioactive BDNF expression, evaluating the effects of medium type (2SY and TM), temperature (25 and 30 °C), and culture time (48–120 h). Maximal BDNF bioactivity (per unit mass) was observed in cultures grown in 2SY medium at extended times (96 h at 30 °C or >72 h at 25 °C), with resulting bioactivity comparable to that of a commercially available BDNF. For cultures grown in 2SY medium at 25 °C for 72 h, the condition that led to the greatest quantity of biologically active protein in the shortest culture time, we recovered 264 μg/L of BDNF. As with other microbial expression systems, BDNF aggregates did form in all culture conditions, indicating that while we were able to recover biologically active BDNF, further optimization of the expression system could yield still greater quantities of bioactive protein. This study provides confirmation that B. choshinensis is capable of producing biologically active BDNF and that further optimization of culture conditions could prove valuable in increasing BDNF yields.
- 11 July 2016
- Published by SAGE Publications in Nucleic Acid Therapeutics
- Vol. 26 (5) , 309-317
Terminal Duplex Stability and Nucleotide Identity Differentially Control siRNA Loading and Activity in RNA Interference
Efficient short interfering RNA (siRNA)-mediated gene silencing requires selection of a sequence that is complementary to the intended target and possesses sequence and structural features that encourage favorable functional interactions with the RNA interference (RNAi) pathway proteins. In this study, we investigated how terminal sequence and structural characteristics of siRNAs contribute to siRNA strand loading and silencing activity and how these characteristics ultimately result in a functionally asymmetric duplex in cultured HeLa cells. Our results reiterate that the most important characteristic in determining siRNA activity is the 5′ terminal nucleotide identity. Our findings further suggest that siRNA loading is controlled principally by the hybridization stability of the 5′ terminus (Nucleotides: 1–2) of each siRNA strand, independent of the opposing terminus. Postloading, RNA-induced silencing complex (RISC)–specific activity was found to be improved by lower hybridization stability in the 5′ terminus (Nucleotides: 3–4) of the loaded siRNA strand and greater hybridization stability toward the 3′ terminus (Nucleotides: 17–18). Concomitantly, specific recognition of the 5′ terminal nucleotide sequence by human Argonaute 2 (Ago2) improves RISC half-life. These findings indicate that careful selection of siRNA sequences can maximize both the loading and the specific activity of the intended guide strand.
- 11 July 2016
- Published by SAGE Publications in Nucleic Acid Therapeutics
- Vol. 26 (5) , 309-317
Terminal Duplex Stability and Nucleotide Identity Differentially Control siRNA Loading and Activity in RNA Interference
Efficient short interfering RNA (siRNA)-mediated gene silencing requires selection of a sequence that is complementary to the intended target and possesses sequence and structural features that encourage favorable functional interactions with the RNA interference (RNAi) pathway proteins. In this study, we investigated how terminal sequence and structural characteristics of siRNAs contribute to siRNA strand loading and silencing activity and how these characteristics ultimately result in a functionally asymmetric duplex in cultured HeLa cells. Our results reiterate that the most important characteristic in determining siRNA activity is the 5′ terminal nucleotide identity. Our findings further suggest that siRNA loading is controlled principally by the hybridization stability of the 5′ terminus (Nucleotides: 1–2) of each siRNA strand, independent of the opposing terminus. Postloading, RNA-induced silencing complex (RISC)–specific activity was found to be improved by lower hybridization stability in the 5′ terminus (Nucleotides: 3–4) of the loaded siRNA strand and greater hybridization stability toward the 3′ terminus (Nucleotides: 17–18). Concomitantly, specific recognition of the 5′ terminal nucleotide sequence by human Argonaute 2 (Ago2) improves RISC half-life. These findings indicate that careful selection of siRNA sequences can maximize both the loading and the specific activity of the intended guide strand.