An acetyltranferase moonlights as a regulator of the RNA binding repertoire of the RNA chaperone Hfq in Escherichia coli.

Bacterial small RNAs (sRNAs) regulate gene expression by base-pairing with their target mRNAs. In Escherichia coli and many other bacteria, this process is dependent on the RNA chaperone Hfq, a mediator for sRNA-mRNA annealing. YhbS (renamed here as HqbA), a putative Gcn5-related N-acetyltransferase (GNAT), was previously identified as a silencer of sRNA signaling in a genomic library screen. Here, we studied how HqbA regulates sRNA signaling and investigated its physiological roles in modulating Hfq activity. Using fluorescent reporter assays, we found that HqbA overproduction suppressed all tested Hfq-dependent sRNA signaling. Direct interaction between HqbA and Hfq was demonstrated both in vivo and in vitro, and mutants that blocked the interaction interfered with HqbA suppression of Hfq. However, an acetylation-deficient HqbA mutant still disrupted sRNA signaling, and HqbA interacted with Hfq at a site far from the active site. This suggests that HqbA may be bifunctional, with separate roles for regulating via Hfq interaction and for acetylation of undefined substrates. Gel shift assays revealed that HqbA strongly reduced the interaction between the Hfq distal face and low-affinity RNAs but not high-affinity RNAs. Comparative RNA immunoprecipitation of Hfq and sequencing showed enrichment of two tRNA precursors, metZWV and proM, by Hfq in mutants that lost the HqbA-Hfq interaction. Our results suggest that HqbA provides a level of quality control for Hfq by competing with low-affinity RNA binders.

Anti-mesothelin immunotoxin induces mesothelioma eradication, anti-tumor immunity, and the development of tertiary lymphoid structures.

LMB-100 is a recombinant immunotoxin composed of a Fab linked to a toxin. It kills cells expressing human mesothelin (hMSLN), which is highly expressed on the surface of mesothelioma and many other cancer cells. Clinically, we observed some patients had delayed responses to an anti-hMSLN immunotoxin treatment, suggesting the induction of anti-tumor immunity. We aimed to develop a mouse model to investigate whether immunotoxin alone can induce anti-tumor immunity and to study the mechanism of this immunity. An immunocompetent transgenic mouse was used to grow mouse mesothelioma AB1 cells expressing hMSLN in the peritoneal cavity. Mice were treated with LMB-100, and mice with complete responses (CRs) were rechallenged with tumor cells to determine whether anti-tumor immunity developed. Changes in gene expression profiles were evaluated by Nanostring, and changes in cytokines and chemokines were checked by protein arrays. The distribution of various immune cells was assessed by immunohistochemistry. Our results show that the mice with tumor reached CRs and developed anti-tumor immunity after LMB-100 treatment alone. The primary response requires CD8+ T cells, CD4+ T cells, and B cells. Transcriptional profiling shows that LMB-100 treatment reshapes the tumor immune microenvironment by upregulating chemotaxis signals. LMB-100 treatment upregulates genes associated with tertiary lymphoid structures (TLS) development and induces TLS formation in tumors. In sum, immunotoxin-mediated cell death induces anti-tumor immunity and the development of TLS, which provides insights into how immunotoxins cause tumor regressions.

Multiple in vivo roles for the C-terminal domain of the RNA chaperone Hfq.

Hfq, a bacterial RNA chaperone, stabilizes small regulatory RNAs (sRNAs) and facilitates sRNA base-pairing with target mRNAs. Hfq has a conserved N-terminal domain and a poorly conserved disordered C-terminal domain (CTD). In a transcriptome-wide examination of the effects of a chromosomal CTD deletion (Hfq1-65), the Escherichia coli mutant was most defective for the accumulation of sRNAs that bind the proximal and distal faces of Hfq (Class II sRNAs), but other sRNAs also were affected. There were only modest effects on the levels of mRNAs, suggesting little disruption of sRNA-dependent regulation. However, cells expressing Hfq lacking the CTD in combination with a weak distal face mutation were defective for the function of the Class II sRNA ChiX and repression of mutS, both dependent upon distal face RNA binding. Loss of the region between amino acids 66-72 was critical for this defect. The CTD region beyond amino acid 72 was not necessary for distal face-dependent regulation, but was needed for functions associated with the Hfq rim, seen most clearly in combination with a rim mutant. Our results suggest that the C-terminus collaborates in various ways with different binding faces of Hfq, leading to distinct outcomes for individual sRNAs.

A fluorescence-based genetic screen reveals diverse mechanisms silencing small RNA signaling in E. coli.

As key players of gene regulation in many bacteria, small regulatory RNAs (sRNAs) associated with the RNA chaperone Hfq shape numerous phenotypic traits, including metabolism, stress response and adaptation, as well as virulence. sRNAs can alter target messenger RNA (mRNA) translation and stability via base pairing. sRNA synthesis is generally under tight transcriptional regulation, but other levels of regulation of sRNA signaling are less well understood. Here we used a fluorescence-based functional screen to identify regulators that can quench sRNA signaling of the iron-responsive sRNA RyhB in Escherichia coli The identified regulators fell into two classes, general regulators (affecting signaling by many sRNAs) and RyhB-specific regulators; we focused on the specific ones here. General regulators include three Hfq-interacting sRNAs, CyaR, ChiX, and McaS, previously found to act through Hfq competition, RNase T, a 3' to 5' exonuclease not previously implicated in sRNA degradation, and YhbS, a putative GCN5-related N-acetyltransferase (GNAT). Two specific regulators were identified. AspX, a 3'end-derived small RNA, specifically represses RyhB signaling via an RNA sponging mechanism. YicC, a previously uncharacterized but widely conserved protein, triggers rapid RyhB degradation via collaboration with the exoribonuclease PNPase. These findings greatly expand our knowledge of regulation of bacterial sRNA signaling and suggest complex regulatory networks for controlling iron homeostasis in bacteria. The fluorescence-based genetic screen system described here is a powerful tool expected to accelerate the discovery of novel regulators of sRNA signaling in many bacteria.

Recombinant immunotoxins with albumin-binding domains have long half-lives and high antitumor activity.

Recombinant immunotoxins (RITs) are chimeric proteins consisting of a Fv that binds to a cancer cell and a portion of a protein toxin. One of these, Moxetumomab pasudotox, was shown to be effective in treating patients with some leukemias, where the cells are readily accessible to the RIT. However, their short half-life limits their efficacy in solid tumors, because penetration into the tumors is slow. Albumin and agents bound to albumin have a long half-life in the circulation. To increase the time tumor cells are exposed to RITs, we have produced and evaluated variants that contain either an albumin-binding domain (ABD) from Streptococcus or single-domain antibodies from Llama. We have inserted these ABDs into RITs targeting mesothelin, between the Fv and the furin cleavage site. We find that these proteins can be produced in large amounts, are very cytotoxic to mesothelin-expressing cancer cell lines, and have a high affinity for human or mouse serum albumin. In mice, the RIT containing an ABD from Streptococcus has a longer half-life and higher antitumor activity than the other two. Its half-life in the circulation of mice ranges from 113 to 194 min compared with 13 min for an RIT with no ABD. Cell uptake studies show the RIT enters the target cell bound to serum albumin. We conclude that RITs with improved half-lives and antitumor activity should be evaluated for the treatment of cancer in humans.

5-Azacytidine prevents relapse and produces long-term complete remissions in leukemia xenografts treated with Moxetumomab pasudotox.

Moxetumomab pasudotox (Moxe) is a chimeric protein composed of an anti-CD22 Fv fused to a portion of Pseudomonas exotoxin A and kills CD22-expressing leukemia cells. It is very active in hairy-cell leukemia, but many children with relapsed/refractory acute lymphoblastic leukemia (ALL) either respond transiently or are initially resistant. Resistance to Moxe in cultured cells is due to low expression of diphthamide genes (DPH), but only two of six ALL blast samples from resistant patients had low DPH expression. To develop a more clinically relevant model of resistance, we treated NSG mice bearing KOPN-8 or Reh cells with Moxe. More than 99.9% of the cancer cells were killed by Moxe, but relapse occurred from discrete bone marrow sites. The resistant cells would no longer grow in cell culture and showed major chromosomal changes and changes in phenotype with greatly decreased CD22. RNA deep sequencing of resistant KOPN-8 blasts revealed global changes in gene expression, indicating dedifferentiation toward less-mature B cell precursors, and showed an up-regulation of myeloid genes. When Moxe was combined with 5-azacytidine, resistance was prevented and survival increased to over 5 months in the KOPN-8 model and greatly improved in the Reh model. We conclude that Moxe resistance in mice is due to a new mechanism that could not be observed using cultured cells and is prevented by treatment with 5-azacytidine.

SymD webserver: a platform for detecting internally symmetric protein structures.

Internal symmetry of a protein structure is the pseudo-symmetry that a single protein chain sometimes exhibits. This is in contrast to the symmetry with which monomers are arranged in many multimeric protein complexes. SymD is a program that detects proteins with internal symmetry. It proved to be useful for analyzing protein structure, function and modeling. This web-based interactive tool was developed by implementing the SymD algorithm. To the best of our knowledge, SymD webserver is the first tool of its kind with which users can easily study the symmetry of the protein they are interested in by uploading the structure or retrieving it from databases. It uses the Galaxy platform to take advantage of its extensibility and displays the symmetry properties, the symmetry axis and the sequence alignment of the structures before and after the symmetry transformation via an interactive graphical visualization environment in any modern web browser. An Example Run video displays the workflow to help users navigate. SymD webserver is publicly available at http://symd.nci.nih.gov.

Assessment of template-free modeling in CASP10 and ROLL.

We present the assessment of predictions for Template-Free Modeling in CASP10 and a report on the first ROLL experiment wherein predictions are collected year round for review at the regular CASP season. Models were first clustered so that duplicated or very similar ones were grouped together and represented by one model in the cluster. The representatives were then compared with targets using GDT_TS, QCS, and three additional superposition-independent score functions newly developed for CASP10. For each target, the top 15 representatives by each score were pooled to form the Top15Union set. All models in this set were visually inspected by four of us independently using the new plugin, EvalScore, which we developed with the UCSF Chimera group. The best models were selected for each target after extensive debate among the four examiners. Groups were ranked by the number of targets (hits) for which a group's model was selected as one of the best models. The Keasar group had most hits in both categories, with four of 19 FM and eight of 36 ROLL targets. The most successful prediction servers were QUARK from Zhang's group for FM category with three hits and Zhang-server for the ROLL category with seven hits. As observed in CASP9, many successful groups were not true "template-free" modelers but used remote templates and/or server models to obtain their winning models. The results of the first ROLL experiment were broadly similar to those of the CASP10 FM exercise.

Assessment of CASP10 contact-assisted predictions.

In CASP10, for the first time, contact-assisted structure predictions have been assessed. Sets of pairs of contacting residues from target structures were provided to predictors for a second round of prediction after the initial round in which they were given only sequences. The objective of the experiment was to measure model quality improvement resulting from the added contact information and thereby assess and help develop so-called hybrid prediction methods--methods where some experimentally determined distance constraints are used to augment de novo computational prediction methods. The results of the experiment were, overall, quite promising.

Definition and classification of evaluation units for CASP10.

For the 10th experiment on Critical Assessment of the techniques of protein Structure Prediction (CASP), the prediction target proteins were broken into independent evaluation units (EUs), which were then classified into template-based modeling (TBM) or free modeling (FM) categories. We describe here how the EUs were defined and classified, what issues arose in the process, and how we resolved them. EUs are frequently not the whole target proteins but the constituting structural domains. However, the assessors from CASP7 on combined more than one domain into 1 EU for some targets, which implied that the assessment also included evaluation of the prediction of the relative position and orientation of these domains. In CASP10, we followed and expanded this notion by defining multidomain EUs for a number of targets. These included 3 EUs, each made of two domains of familiar fold but arranged in a novel manner and for which the focus of evaluation was the interdomain arrangement. An EU was classified to the TBM category if a template could be found by sequence similarity searches and to FM if a structural template could not be found by structural similarity searches. The EUs that did not fall cleanly in either of these cases were classified case-by-case, often including consideration of the overall quality and characteristics of the predictions.

Discovering Novel Bacterial Small RNA by RNA-seq Analysis Toolkit ANNOgesic.

ANNOgesic is an RNA-seq analysis pipeline that can detect sRNAs and many other genomic features in bacteria and archaea. In addition to listing sRNA candidates, ANNOgesic also generates various formats of data files for visual examination and downstream experimental design. Based on validations from previous studies, the sRNA predictions are accurate and reliable. In this chapter, we outline the sRNA detection algorithm, important parameters used, step-by-step execution, and data interpretation with a B. pertussis study as an example. Following those procedures, novel sRNA can be revealed by ANNOgesic.

DOMIRE: a web server for identifying structural domains and their neighbors in proteins.

SUMMARY: The DOMIRE web server implements a novel, automatic, protein structural domain assignment procedure based on 3D substructures of the query protein which are also found within structures of a non-redundant protein database. These common 3D substructures are transformed into a co-occurrence matrix that offers a global view of the protein domain organization. Three different algorithms are employed to define structural domain boundaries from this co-occurrence matrix. For each query, a list of structural neighbors and their alignments are provided. DOMIRE, by displaying the protein structural domain organization, can be a useful tool for defining protein common cores and for unravelling the evolutionary relationship between different proteins. AVAILABILITY: http://genome.jouy.inra.fr/domire CONTACT: jean.garnier@jouy.inra.fr.

Camel nanobody-based B7-H3 CAR-T cells show high efficacy against large solid tumours.

Rational design of chimeric antigen receptor T (CAR-T) cells based on the recognition of antigenic epitopes capable of evoking the most potent CAR activation is an important objective in optimizing immune therapy. In solid tumors, the B7-H3 transmembrane protein is an emerging target that harbours two distinct epitope motifs, IgC and IgV, in its ectodomain. Here, we generate dromedary camel nanobodies targeting B7-H3 and demonstrate that CAR-T cells, based on the nanobodies recognizing the IgC but not IgV domain, had potent antitumour activity against large tumors in female mice. These CAR-T cells are characterized by highly activated T cell signaling and significant tumor infiltration. Single-cell transcriptome RNA sequencing coupled with functional T-cell proteomics analysis uncovers the top-upregulated genes that might be critical for the persistence of polyfunctional CAR-T cells in mice. Our results highlight the importance of the specific target antigen epitope in governing optimal CAR-T activity and provide a nanobody-based B7-H3 CAR-T product for use in solid tumor therapy.

Protein coopted from a phage restriction system dictates orthogonal cell division plane selection in Staphylococcus aureus.

The spherical bacterium Staphylococcus aureus, a leading cause of nosocomial infections, undergoes binary fission by dividing in two alternating orthogonal planes, but the mechanism by which S. aureus correctly selects the next cell division plane is not known. To identify cell division placement factors, we performed a chemical genetic screen that revealed a gene which we termed pcdA. We show that PcdA is a member of the McrB family of AAA+ NTPases that has undergone structural changes and a concomitant functional shift from a restriction enzyme subunit to an early cell division protein. PcdA directly interacts with the tubulin-like central divisome component FtsZ and localizes to future cell division sites before membrane invagination initiates. This parallels the action of another McrB family protein, CTTNBP2, which stabilizes microtubules in animals. We show that PcdA also interacts with the structural protein DivIVA and propose that the DivIVA/PcdA complex recruits unpolymerized FtsZ to assemble along the proper cell division plane. Deletion of pcdA conferred abnormal, non-orthogonal division plane selection, increased sensitivity to cell wall-targeting antibiotics, and reduced virulence in a murine infection model. Targeting PcdA could therefore highlight a treatment strategy for combatting antibiotic-resistant strains of S. aureus.

iCn3D: From Web-Based 3D Viewer to Structural Analysis Tool in Batch Mode.

iCn3D was initially developed as a web-based 3D molecular viewer. It then evolved from visualization into a full-featured interactive structural analysis software. It became a collaborative research instrument through the sharing of permanent, shortened URLs that encapsulate not only annotated visual molecular scenes, but also all underlying data and analysis scripts in a FAIR manner. More recently, with the growth of structural databases, the need to analyze large structural datasets systematically led us to use Python scripts and convert the code to be used in Node. js scripts. We showed a few examples of Python scripts at https://github.com/ncbi/icn3d/tree/master/icn3dpython to export secondary structures or PNG images from iCn3D. Users just need to replace the URL in the Python scripts to export other annotations from iCn3D. Furthermore, any interactive iCn3D feature can be converted into a Node. js script to be run in batch mode, enabling an interactive analysis performed on one or a handful of protein complexes to be scaled up to analysis features of large ensembles of structures. Currently available Node. js analysis scripts examples are available at https://github.com/ncbi/icn3d/tree/master/icn3dnode. This development will enable ensemble analyses on growing structural databases such as AlphaFold or RoseTTAFold on one hand and Electron Microscopy on the other. In this paper, we also review new features such as DelPhi electrostatic potential, 3D view of mutations, alignment of multiple chains, assembly of multiple structures by realignment, dynamic symmetry calculation, 2D cartoons at different levels, interactive contact maps, and use of iCn3D in Jupyter Notebook as described at https://pypi.org/project/icn3dpy.

Protein domain assignment from the recurrence of locally similar structures.

Domains are basic units of protein structure and essential for exploring protein fold space and structure evolution. With the structural genomics initiative, the number of protein structures in the Protein Databank (PDB) is increasing dramatically and domain assignments need to be done automatically. Most existing structural domain assignment programs define domains using the compactness of the domains and/or the number and strength of intra-domain versus inter-domain contacts. Here we present a different approach based on the recurrence of locally similar structural pieces (LSSPs) found by one-against-all structure comparisons with a dataset of 6373 protein chains from the PDB. Residues of the query protein are clustered using LSSPs via three different procedures to define domains. This approach gives results that are comparable to several existing programs that use geometrical and other structural information explicitly. Remarkably, most of the proteins that contribute the LSSPs defining a domain do not themselves contain the domain of interest. This study shows that domains can be defined by a collection of relatively small locally similar structural pieces containing, on average, four secondary structure elements. In addition, it indicates that domains are indeed made of recurrent small structural pieces that are used to build protein structures of many different folds as suggested by recent studies.

Overexpression of the Bacteriophage T4 motB Gene Alters H-NS Dependent Repression of Specific Host DNA.

The bacteriophage T4 early gene product MotB binds tightly but nonspecifically to DNA, copurifies with the host Nucleoid Associated Protein (NAP) H-NS in the presence of DNA and improves T4 fitness. However, the T4 transcriptome is not significantly affected by a motB knockdown. Here we have investigated the phylogeny of MotB and its predicted domains, how MotB and H-NS together interact with DNA, and how heterologous overexpression of motB impacts host gene expression. We find that motB is highly conserved among Tevenvirinae. Although the MotB sequence has no homology to proteins of known function, predicted structure homology searches suggest that MotB is composed of an N-terminal Kyprides-Onzonis-Woese (KOW) motif and a C-terminal DNA-binding domain of oligonucleotide/oligosaccharide (OB)-fold; either of which could provide MotB's ability to bind DNA. DNase I footprinting demonstrates that MotB dramatically alters the interaction of H-NS with DNA in vitro. RNA-seq analyses indicate that expression of plasmid-borne motB up-regulates 75 host genes; no host genes are down-regulated. Approximately 1/3 of the up-regulated genes have previously been shown to be part of the H-NS regulon. Our results indicate that MotB provides a conserved function for Tevenvirinae and suggest a model in which MotB functions to alter the host transcriptome, possibly by changing the association of H-NS with the host DNA, which then leads to conditions that are more favorable for infection.

Identification of BvgA-Dependent and BvgA-Independent Small RNAs (sRNAs) in Bordetella pertussis Using the Prokaryotic sRNA Prediction Toolkit ANNOgesic.

Noncoding small RNAs (sRNAs) are crucial for the posttranscriptional regulation of gene expression in all organisms and are known to be involved in the regulation of bacterial virulence. In the human pathogen Bordetella pertussis, which causes whooping cough, virulence is controlled primarily by the master two-component system BvgA (response regulator)/BvgS (sensor kinase). In this system, BvgA is phosphorylated (Bvg+ mode) or nonphosphorylated (Bvg- mode), with global transcriptional differences between the two. B. pertussis also carries the bacterial sRNA chaperone Hfq, which has previously been shown to be required for virulence. Here, we conducted transcriptomic analyses to identify possible B. pertussis sRNAs and to determine their BvgAS dependence using transcriptome sequencing (RNA-seq) and the prokaryotic sRNA prediction program ANNOgesic. We identified 143 possible candidates (25 Bvg+ mode specific and 53 Bvg- mode specific), of which 90 were previously unreported. Northern blot analyses confirmed all of the 10 ANNOgesic candidates that we tested. Homology searches demonstrated that 9 of the confirmed sRNAs are highly conserved among B. pertussis, Bordetella parapertussis, and Bordetella bronchiseptica, with one that also has homologues in other species of the Alcaligenaceae family. Using coimmunoprecipitation with a B. pertussis FLAG-tagged Hfq, we demonstrated that 3 of the sRNAs interact directly with Hfq, which is the first identification of sRNA binding to B. pertussis Hfq. Our study demonstrates that ANNOgesic is a highly useful tool for the identification of sRNAs in this system and that its combination with molecular techniques is a successful way to identify various BvgAS-dependent and Hfq-binding sRNAs. IMPORTANCE Noncoding small RNAs (sRNAs) are crucial for posttranscriptional regulation of gene expression in all organisms and are known to be involved in the regulation of bacterial virulence. We have investigated the presence of sRNAs in the obligate human pathogen B. pertussis, using transcriptome sequencing (RNA-seq) and the recently developed prokaryotic sRNA search program ANNOgesic. This analysis has identified 143 sRNA candidates (90 previously unreported). We have classified their dependence on the B. pertussis two-component system required for virulence, namely, BvgAS, based on their expression in the presence/absence of the phosphorylated response regulator BvgA, confirmed several by Northern analyses, and demonstrated that 3 bind directly to B. pertussis Hfq, the RNA chaperone involved in mediating sRNA effects. Our study demonstrates the utility of combining RNA-seq, ANNOgesic, and molecular techniques to identify various BvgAS-dependent and Hfq-binding sRNAs, which may unveil the roles of sRNAs in pertussis pathogenesis.

Long term follow-up of a phase II study of cladribine with concurrent rituximab with hairy cell leukemia variant.

Hairy cell leukemia variant (HCLv) responds poorly to purine analogue monotherapy. Rituximab concurrent with cladribine (CDAR) improves response rates, but long-term outcomes are unknown. We report final results of a phase 2 study of CDAR for patients with HCLv. Twenty patients with 0 to 1 prior courses of cladribine and/or rituximab, including 8 who were previously untreated, received cladribine 0.15 mg/kg on days 1 to 5 with 8 weekly rituximab doses of 375 mg/m2 beginning day 1. Patients received a second rituximab course ≥6 months after cladribine, if and when minimal residual disease (MRD) was detected in blood. The complete remission (CR) rate from CDAR was 95% (95% confidence interval, 75-100). Sixteen (80%) of 20 patients (95% confidence interval, 56-94) became MRD negative according to bone marrow at 6 months. The median duration of MRD-negative CR was 70.1 months, and 7 of 16 are still MRD negative up to 120 months. With a median follow-up of 69.7 months, 11 patients received delayed rituximab, and the 5-year progression-free survival (PFS) and overall survival (OS) were 63.3% and 73.9%, respectively. Five patients with TP53 mutations had shorter PFS (median, 36.4 months vs unreached; P = .0024) and OS (median, 52.4 months vs unreached; P = .032). MRD-negative CR at 6 months was significantly associated with longer PFS (unreached vs 17.4 months; P < .0001) and OS (unreached vs 38.2 months; P < .0001). Lack of MRD in blood at 6 months was also predictive of longer PFS and OS (P < .0001). After progression following CDAR, median OS was 29.7 months. CDAR is effective in HCLv, with better outcomes in patients who achieve MRD-negative CR. This trial is registered at www.clinicaltrials.gov as #NCT00923013.

CAR T cells targeting tumor-associated exons of glypican 2 regress neuroblastoma in mice.

Targeting solid tumors must overcome several major obstacles, in particular, the identification of elusive tumor-specific antigens. Here, we devise a strategy to help identify tumor-specific epitopes. Glypican 2 (GPC2) is overexpressed in neuroblastoma. Using RNA sequencing (RNA-seq) analysis, we show that exon 3 and exons 7-10 of GPC2 are expressed in cancer but are minimally expressed in normal tissues. Accordingly, we discover a monoclonal antibody (CT3) that binds exons 3 and 10 and visualize the complex structure of CT3 and GPC2 by electron microscopy. The potential of this approach is exemplified by designing CT3-derived chimeric antigen receptor (CAR) T cells that regress neuroblastoma in mice. Genomic sequencing of T cells recovered from mice reveals the CAR integration sites that may contribute to CAR T cell proliferation and persistence. These studies demonstrate how RNA-seq data can be exploited to help identify tumor-associated exons that can be targeted by CAR T cell therapies.