Small-molecule drug discovery relies on identifying compounds that modulate specific protein targets, a process often hindered by cellular complexity. Through phenotypic screening of a kinase-focused diazaquinazoline library, we serendipitously identified CEM198 as the first high-affinity ligand of tubulin-tyrosine ligase (TTL). Functional assays combining live-cell TTL inhibition, microtubule polymerization, cell cycle analysis, and proteomics revealed that CEM198 acts through a dual mechanism: directly binding to TTL and altering α/β-tubulin conformation. This interaction restricts α-tubulin tyrosination and disrupts tubulin polymerization, leading to microtubule destabilization. The differential effects observed between SH-SY5Y and HEK293T cells indicate that effective TTL inhibition depends on both direct binding and structural modulation of the tubulin heterodimer. These findings introduce CEM198 as a chemical probe for investigating the tubulin tyrosination - detyrosination and demonstrate the potential of chemoproteomics to uncover novel modulators of microtubule dynamics.

When organisms encounter pathogens, they rapidly activate complex defense programs to ensure survival. While these immune responses are vital, they often also incur trade-offs, such as reduced growth and development and must therefore be tightly controlled. In this study, we reveal that the steroid hormones brassinosteroids (BRs) contribute to this control in Arabidopsis thaliana by repressing immunity-related genes. We provide evidence that the BR-regulated basic helix-loop-helix (bHLH) transcription factor CESTA (CES), along with its homologs BR ENHANCED EXPRESSION (BEE)1-3, mediate DNA methylation changes at transposable element (TE)-rich loci containing nucleotide-binding leucine-rich-repeat (NLR)-type receptor genes, including SUPPRESSOR OF NPR1-1 CONSTITUTIVE 1 (SNC1). These CES-induced methylation changes correlate with altered splicing of SNC1 pre-mRNA, a process that requires the BR receptor BRASSINOSTEROID INSENSITIVE 1 (BRI1). In support, we show that CES associates with components of the chromatin remodeling and splicing machinery. Together, our findings reveal a previously unrecognized BR-induced mechanism that modulates the epigenetic and post transcriptional regulation of immune genes, enabling plants to prioritize growth over defense.

Maize is a globally important staple crop, providing food, animal feed, and industrial raw materials. Its extensive native diversity can be utilized to broaden the genetic basis for quantitative trait improvement and advance our understanding of the genotype-phenotype relationships underlying complex trait variation. In a forward genetic approach, we dissect a locus on chromosome 10 with major effects on Photosystem II (PSII) maximum quantum efficiency (Fv/Fm), non-photochemical quenching (NPQ) and biomass accumulation during early growth stages. Integrating molecular and physiological information we show that allelic variation at the gene encoding LIGHT HARVESTING CHLOROPHYLL A/B BINDING PROTEIN6 (LHCB6 also known as CP24), a component of the PSII light-harvesting complex (LHCII) antenna, underlies this locus. We demonstrate that the allelic variation results from a hAT transposon insertion at lhcb6 and is associated with decreased accumulation of the LHCII antenna components LHCB6 and LHCB3. Based on proteomic analyses we propose candidate genes that partially compensate for the unfavorable early growth effects caused by impaired LHCII antenna assembly. Our work provides novel insights into the function of lhcb6 in the C4 crop maize and demonstrates the value of natural variation for the understanding and genetic improvement of complex photosynthetic processes.

Abstract YAP and TAZ are Hippo pathway effectors that bind TEAD transcription factors to drive oncogenic programs. First-in-human trials of TEAD inhibitors (TEADis) showed activity and tolerability in pretreated mesothelioma patients, yet current trials focus on mesotheliomas and rare Hippo-altered tumors, despite the broad oncogenic role of YAP/TAZ in various solid cancers. Whether YAP/TAZ target gene signatures can identify tumors with high YAP/TAZ activity and potential TEADi sensitivity beyond these indications remains unclear. We profiled YAP/TAZ activity using a 22-gene signature across 2390 advanced pediatric and 3746 advanced rare cancers or cancers from young adults (<51y) from the INFORM and DKFZ/NCT/DKTK MASTER (MASTER) trials. Benchmarking against mesothelioma and YAP fusion-driven ependymoma revealed subsets of cancers across entity baskets with equal or higher YAP/TAZ activity independent of Hippo alterations. To dissect tumor-intrinsic versus stromal contributions, we performed spatial transcriptomics on MASTER samples using microarrays and the Xenium platform (n=106 cores from 89 tumors representing >70 histological subtypes). Tumor-cell YAP/TAZ scores correlated strongly with bulk scores (r=0.61), whereas fibroblast (r=0.23) and immune-cell (r=0.35) correlations were weak, indicating that bulk scores largely reflect tumor-cell activity. We highlight a MASTER patient with MET-amplified carcinoma in whom progression on the MET inhibitor capmatinib coincided with a marked YAP/TAZ-score increase. In MET-amplified HS746T gastric cancer cells, genetic YAP/TAZ activation reduced capmatinib sensitivity, partially reversed by TEAD inhibition, supporting an unrecognized YAP/TAZ-linked resistance mechanism. To evaluate predictive capacity of the YAP/TAZ signature, 23 cancer cell lines and 30 patient-derived spheroid cultures (PDSCs) were treated with IAG933 (Ω-loop-binding TEADi) or VT107 (TEAD autopalmitoylation inhibitor). Across compounds and models, expression of genes most strongly associated with TEADi response were enriched for YAP/TAZ targets, enabling refinement of a TEADi response signature. This refined signature correlated with area-under-the-drug-response-curve (AUC) values (IAG933: Pearson r=0.81 for cell lines, r=0.60 for PDSCs; VT107: r=0.46 and r=0.45). Notably, the refined signature derived from INFORM/MASTER tumors predicted TEADi response in matched PDSCs (n=13; r=0.70 for IAG933, r=0.72 for VT107), with IC50 values as low as 30 nM. Sensitive PDSCs included atypical rhabdoid tumor, colorectal, and pancreatic carcinoma - all lacking Hippo alterations and representing entities not typically considered YAP/TAZ-driven. In summary, these data provide a strong rationale for a transcription-based stratified clinical trial evaluating TEAD inhibition across pediatric and adult advanced cancers irrespective of Hippo alterations. Citation Format: Michael Wegert-Verhoeven, Mathea Fransisca, Sylvia Martin, Attila Jady, Daniela Richter, Malgorzata Oles, Timon A. Blindauer, Jan-Philipp Mallm, Jasmina Paluncic, Claudia Dagostino, Olga Ermakova, Annika Schneider, Matthew The, Annika Baude-Müller, Katja Beck, Maximilian Bullemer, Victor Didier Meza, Vivek Venkataramani, Ralf C. Bargou, Heiko Becker, Melanie Boerries, Armin Tuchscherer, DKFZ/NCT/DKTK MASTER consortium, Martin Wermke, Andreas Brunschweiger, Mohammed Al-Saeedi, Dirk Jäger, Olaf Witt, Denis Schapiro, Bernhard Küster, Andreas Hartig, Michael Allgaeuer, Alexander Brobeil, Christoph E. Heilig, Maria Veronica Teleanu, Simon Kreutzfeldt, Peter Horak, Daniel Hübschmann, Wolfgang Hartmann, Marcel Trautmann, Ina Oehme, Claudia R. Ball, Stefan Fröhling, Stefan M. Pfister, Hanno Glimm, Sebastian M. Dieter. A tumor-agnostic YAP/TAZ score predicts TEAD inhibitor sensitivity independent of Hippo alterations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 1028.

Abstract The molecular tumor board (MTB) is central to precision oncology, providing personalized treatment recommendations based on molecular profiles of patient tumors. Genomics is instrumental for MTBs but often fails to identify clinically actionable targets, a gap that phosphoproteomics can fill. We present the tumor proteome activity status (TOPAS) platform, an end-to-end analysis pipeline that converts terabytes of phosphoproteomic data into patient-specific reports for MTB discussions, focusing on clinically relevant signaling linked to oncogenic mechanisms and therapeutic targets. Designed to scale with growing cohorts, the platform integrates data from 1,998 tumor samples to support patient- and cohort-level hypothesis generation. A web portal handles quality control, calculates TOPAS scores, identifies tumor antigens and immune checkpoints, and offers interactive analyses of differential protein abundance and outlier detection. The TOPAS platform is open source, addresses a critical unmet need and facilitates broader adoption of phosphoproteomics in precision oncology in the future.

Citrullination is a post-translational modification implicated in autoimmune and inflammatory diseases, yet its low abundance and lack of effective enrichment tools have limited proteome-wide analysis. Here, we develop a robust chemical proteomics workflow with improved specificity and throughput. This method builds upon glyoxal-based derivatization and incorporates a cleavable biotin linker for efficient peptide enrichment, release, and identification via mass spectrometry. Benchmarking demonstrates a > 10-fold increase in the detection of citrullinated peptides at sub-0.1% abundance. Applying this workflow to primary human neutrophils, we successfully monitor dynamic regulation, quantifying dose-dependent activation and inhibition by the PAD4 inhibitor GSK484. Furthermore, stimulation with the fungal pathogen Candida albicans reveals a "core citrullinome" conserved across distinct stimuli. Notably, extensive citrullination of linker histone H1 and structural proteins like lamin B1 suggests broad remodeling of cell architecture during NET formation. This workflow enables proteome-wide mapping of citrullination sites and facilitates its study across diverse biological contexts.

Proteomic and phenotypic cell sensitivity datasets are increasingly important for understanding chemoproteomics and the underlying drug mechanisms of action. Yet, integrating such heterogeneous datasets remains challenging due to inconsistent annotations, incompatible IDs, and variable data processing methods. Here, a major update to ProteomicsDB (https://www.proteomicsdb.org) is presented that combines over 1300 proteomic and 1000 transcriptomic profiles with phenotypic cell sensitivity data across >1500 human cancer cell lines and 1470 drugs. Harmonizing cell line and drug names and applying a standardized normalization and refitting pipeline for dose–response curves enables consistent, statistically robust analysis across studies. Three new graphical user interfaces support interactive exploration of cell sensitivity data, exploring the protein targets and dose-resolved changes in protein expression in the presence of a drug, and comparing the expression profiles of cell lines. With this update, ProteomicsDB is strengthening its future role as a central hub for proteomics and multi-omics, providing researchers with a unified framework to explore phenotypic cell sensitivity in combination with dose-resolved expression proteomics at the molecular level, supporting biomarker discovery, drug repurposing, and precision medicine applications.

Immune checkpoint inhibitors (ICIs) have markedly improved outcomes in malignant melanoma, yet accessible, tissue-based predictive biomarkers of response remain limited. PD-L1 expression, which remains widely used, offers some prognostic value but performs poorly at low expression levels. We performed liquid chromatography-mass spectrometry proteomic profiling on 185 melanoma samples, including 52 pretreatment samples from ICI responders and nonresponders. Differential expression and pathway analyses identified a 73-protein immune activation signature correlating with ICI response. Within this immune context, SH2D1A (SAP), a regulator of T- and natural killer-cell cytotoxicity, emerged as a top candidate biomarker. Internal immunohistochemical validation confirmed its strong predictive accuracy (area under the receiver operating characteristic curve: 0.93; sensitivity: 88%; specificity: 91%) in distinguishing responders from nonresponders, outperforming PD-L1, CD3, and CD8. Notably, SH2D1A retained its predictive power in PD-L1-low tumors (combined positive score < 10), suggesting clinical use in which the current markers fall short. Independent validation in an external tissue microarray cohort revealed attenuated predictive performance, likely reflecting the impact of intratumoral heterogeneity, as further suggested by whole-slide validation. Nevertheless, SH2D1A remained significantly enriched in tumors that responded to ICIs and was associated with improved survival outcomes, outperforming CD3, CD8, and PD-L1. In conclusion, these results suggest that SH2D1A provides information beyond established immune infiltration- and exhaustion-related markers, supporting further investigation of its potential role in biomarker-guided prediction of ICI response in melanoma.

The activation of chemical reactions in living cells using ultraviolet (UV) light enables the interrogation of biomolecules in their native environment with photoreactive probes or crosslinking reagents. Although numerous photo-crosslinking approaches have been successfully employed, they often suffer from common limitations, including low reaction yields, the need for long exposure times, and irradiation-induced cellular damage from heat, desiccation, or side reactions. We recently showed that 365 nm light-emitting diodes enable rapid, biorthogonal protein-DNA crosslinking in living cells, incurring minimal photodamage. Here, we generalize this approach and demonstrate that high-intensity, longwave UV light reduces the irradiation time for in-cell photo-crosslinking reactions by up to 1000-fold, allowing protein-drug, protein-protein, protein-DNA, and protein-RNA interactions to be fixed within seconds. Benchmarking this rapid photo-activation for the analysis of RNA-interacting proteomes responding to RNA-binding drugs or UV-induced RNA damage, we demonstrate both qualitative and quantitative advantages of controlled, high-intensity UV irradiation, uncovering emergent experimental opportunities that were previously inaccessible to light-activated chemistry in intact cells and tissues.

Plant cells are enclosed by a semi-rigid cell wall with a complex biochemical composition and architecture. The poorly understood process of remodelling the cell wall is crucial for controlling growth and development and for regulating abiotic and biotic stress responses. Cell wall remodelling upon disruption of cell wall integrity through inhibition of cellulose biosynthesis depends on the receptor kinase STRUBBELIG (SUB) and its binding partner QUIRKY (QKY). Here, we identify NON-RACE SPECIFIC DISEASE RESISTANCE/HIN1 HAIRPIN-INDUCED-LIKE protein 3 (NHL3) as an additional factor involved in the SUB-dependent cellulose biosynthesis inhibition response. Collectively, our data indicate that NHL3 maintains SUB at the plasma membrane by physically interacting with SUB. SUB signalling is attenuated by receptor-mediated endocytosis initiated by release of first the SUB-NHL3 and later the SUB-QKY interaction. Our results further suggest a diverse set of biochemically and functionally distinct SUB complexes involved in regulating cell wall integrity and development.

The oncogenic kinase AURORA A is essential for mitotic progression, and its catalytic inhibition arrests cells at the G2/M-transition. Unexpectedly, degradation of AURORA A by PROTACs (proteolysis targeting chimeras) induces profound S-phase defects, revealing a non-catalytic scaffolding function of AURORA A. To dissect this function, we profile the AURORA A S-phase interactome and identify multiple RNA-binding proteins not characterized as AURORA A substrates. Among these, the ribonuclease DICER directly associates with AURORA A to form an abundant nuclear complex. RNA degradation shifts AURORA A, DICER, and additional RNA-binding proteins from heavy to light gradient fractions, implicating RNA-dependent complex function. In contrast, PROTAC-mediated depletion of AURORA A alters the gradient migration behavior and chromatin association of the histone methyltransferase SETD2, which is known to prevent spurious transcription. These findings reveal a dual-output model for the S-phase AURORA A complex: First, RNA-binding proteins are recruited to R-loops, which may arise from transcription-replication conflicts. DICER then processes the R-loop, while AURORA A simultaneously recruits SETD2, which facilitates efficient resolution of replicative stress by preventing spurious transcription.

The clinical and molecular heterogeneity of diffuse large B cell lymphoma (DLBCL) is incompletely understood. By integrating proteomic, transcriptomic, and genomic data from 478 DLBCL tumors, we identify seven DLBCL proteogenotypes (PGs) reflecting specific pathophysiological features that span known molecular subtypes. PG4 is associated with poor outcome independent of established risk factors such as cell-of-origin, international prognostic index, or genetic features. PG4 contains activated B cell-like and germinal center B cell-like tumors and genetically unclassified cases. It shares a dark-zone-related B cell phenotype and shows enrichment for BTG1 mutations that can activate MYC. Single-cell sequencing and spatial transcriptomics reveal enhanced MYC and TCF3/4 transcriptional activity irrespective of MYC translocations. The PG4 tumor microenvironment is characterized by exhausted CD8+ T cells. Our study identifies common oncogenic themes underlying high-risk DLBCL tumors and provides a proteogenomic framework for future diagnostic and therapeutic approaches.

1. Abstract Accurate genome annotation is fundamental to modern biology, yet distinguishing authentic protein-coding sequences from prediction artifacts remains challenging, particularly in complex plant genomes where automated methods are error-prone and manual curation is rarely feasible due to prohibitive time and costs. Here, we present GAP-MS ( G ene model A ssessment using P eptides from M ass S pectrometry), an automated proteogenomic pipeline that leverages mass spectrometry evidence to validate the protein-level accuracy of predicted gene models. Applied across 9 major crop species, GAP-MS consistently improved the prediction precision for four widely used gene prediction tools. In addition to filtering likely erroneous models, the pipeline identified hundreds of candidate protein-coding loci absent from current standard reference annotations. These peptide-supported loci were further verified by transcriptional evidence, well-supported functional annotations, and high coding-potential scores. Together, these results demonstrate that direct proteomic evidence can help resolve annotation ambiguities, define high-confidence peptide-supported reference proteomes, and uncover overlooked protein-coding genes, while facilitating the identification of sequences that may require further investigation.

BACKGROUND:Proteasome inhibitors (PIs) are indispensable for the treatment of multiple myeloma (MM), the second most common hematologic malignancy. Although primary resistance to PIs is rare, most patients eventually relapse and develop acquired resistance, with underlying mechanisms that remain incompletely understood and appear to be drug-specific. In the case of bortezomib, resistance is often associated with PSMB5 mutations. In contrast, resistance to carfilzomib (CFZ) is mediated by overexpression of the drug efflux transporter ABCB1. However, the regulatory mechanisms driving ABCB1 upregulation in CFZ-resistant MM remain unclear. METHODS:An integrative multi-omics analysis was conducted using paired samples from a CFZ-sensitive and -resistant MM patient, alongside resistant cell line models. Whole-genome sequencing (WGS), whole-genome bisulfite sequencing (WGBS), and RNA sequencing (RNA-seq) were used to assess the genotype (structural variants, single nucleotide variants, and copy number variations), methylation status, and the expression of the ABCB1 locus. ABCB1 promoter methylation levels and expression levels in an independent MM subcohort were analyzed to determine clinical relevance. Functional validation was performed using dual-luciferase reporter assays, DNMT1 knockdown, and treatment with DNA methyltransferase inhibitors (DNMTis) to evaluate methylation-dependent regulation of ABCB1 expression. RESULTS:Significant hypomethylation of the ABCB1 downstream promoter region was identified (GH07J087598) in a CFZ-resistant patient sample, which correlated with elevated ABCB1 expression. Consistent with the paired CFZ-resistant case, the independent MM subcohort showed a significant inverse association between ABCB1 promoter methylation and ABCB1 expression. These findings align with results obtained from CFZ-resistant MM cell line models, which demonstrated reduced promoter methylation and elevated ABCB1 expression compared to their wild-type counterparts. Furthermore, treatment with DNA methyltransferase inhibitors as well as DNMT1 knockdown enhanced ABCB1 expression while demethylating the promoter, thereby validating the functional significance of promoter hypomethylation in ABCB1 overexpression. CONCLUSIONS:Our findings highlight ABCB1 promoter hypomethylation as a potential epigenetic driver of CFZ resistance in MM. These results underscore the clinical relevance of epigenetic regulation in drug resistance and the potential of targeting DNA methylation as a therapeutic strategy to overcome resistance in MM.

Chimeric antigen receptor (CAR) T cells exhibit high response rates in B-cell malignancies, but most patients eventually relapse. A key mechanism of treatment failure is the loss or downregulation of tumor antigen expression, yet strategies to modulate cell surface levels of CAR T-cell targets remain largely unexplored. Here, we identify B-cell maturation antigen (BCMA), a central CAR T-cell target in multiple myeloma (MM), as a highly shortlived protein that undergoes K48-linked polyubiquitylation at the plasma membrane, leading to its p97-dependent degradation via the ubiquitin-proteasome system (UPS). This previously unprecedented mechanism of plasma membrane protein regulation enables significant enhancement of BCMA expression via proteasome inhibitors (PIs). The clinically approved PI carfilzomib (CFZ) significantly enhances the efficacy of BCMAdirected CAR T cells against both PI-sensitive and-refractory MM cells in vitro and in vivo. Notably, CFZ treatment of 10 patients with BMCA CAR T-cell therapy relapse, under the CFZ after BCMA CAR T-cell (CarCAR) protocol, resulted in increased BCMA expression in all patients. However, clinical responses were observed only in those with residual and/or expanding CAR T cells, suggesting restored CAR T-cell function. These findings provide a rationale for the use of CFZ treatment in relapsed or refractory MM after BCMA CAR T-cell therapy, advocate for future trials combining CFZ with BCMA CAR T cells, and provide a framework for exploring UPS-dependent degradation of other immunotherapy antigens.

Genomics-guided precision oncology has improved survival in cancer entities with actionable mutations but cannot capture oncogenic signaling that manifests at the protein level. Here, we report a prospective, real-world pan-cancer study profiling proteomes and phosphoproteomes of 1,998 tumor samples from adults and children with rare or advanced cancers enrolled in the German precision oncology programs DKFZ/NCT/DKTK MASTER, CATCH and INFORM and their molecular tumor boards (MTBs). We developed tumor proteome activity status (TOPAS) scores for 46 clinically relevant kinases, an immune activity score capturing antigen presentation and T-cell activation and identified therapeutically targetable cell-surface proteins for 94% of patients. These readouts enhance MTB recommendations by exposing actionable non-genomic kinase activity, refining interpretation of oncogenic genome alterations, and highlighting cell-surface treatment options. Three proof-of-concept analyses indicate clinical utility including kinase activity-stratified pazopanib response in sarcoma, immune activity score-tracked checkpoint-inhibitor outcomes pan-cancer, and a phosphoproteomic biomarker distinguishing EGFR-inhibitor response in chordoma.

Immunotherapies targeting surface antigens have transformed the treatment landscape of multiple myeloma (MM), with GPRC5D emerging as a promising therapeutic target. Monoallelic loss of GPRC5D is frequently observed in newly diagnosed MM patients, and the incidence of acquired GPRC5D alterations increases following exposure to GPRC5D-directed therapies. However, the functional consequences of both baseline monoallelic and therapy-induced biallelic GPRC5D alterations remain poorly understood. In this study, we modeled monoallelic versus biallelic loss of GPRC5D to investigate their impact on MM cell biology and responsiveness to GPRC5D-targeted immunotherapies. Our results demonstrate that monoallelic GPRC5D loss in OPM-2 cells reduces surface expression of the antigen and confers resistance to GPRC5D-directed therapies. Complete loss of GPRC5D alters the transcriptional state of MM cells and promotes reprogramming of the phosphoproteomic circuitry ultimately resulting in a pro-proliferative chemokine environment. As a result, GPRC5D deficiency increases the basal proliferation rate of MM cells thereby providing a competitive advantage which may further be amplified by selecting these aggressive phenotypes during ongoing treatment with anti-GPRC5D immunotherapies.

The gut microbiota influences host metabolism, but the mechanisms of lipid uptake from food remain mysterious. Here we used stable isotope-labelled tracers in gnotobiotic mouse models, which revealed that host uptake of dietary lipids depends on microbial colonization. Systemic lipid metabolism modelling predicted that the gut microbiota restricts intestinal lipid absorption, and labelled lipid administration verified that the gut contents of microbiota-colonized mice contained up to 12-fold more lipids than those of germ-free animals. A combination of lipidomics and proteomics showed that gut microbes trigger Myd88 signalling, leading to a downregulation of hepatic Cyp7b1 activity and increased taurocholate production. Taurocholate stimulates phospholipase A1 activity in bile, causing the degradation of phosphatidylcholine that is essential for luminal micelle formation and lipid uptake. A diverse microbiome was associated with lower phosphatidylcholine content. This previously unrecognized host-gut microbiota interplay via enzymes in bile could provide future targets to modulate dietary lipid absorption.

In flowering plants, the correct timing of pollen germination is crucial for successful fertilization and, therefore, an important trait for seed and fruit crops. Environmental factors tightly regulate pollen germination, with germination-suppressing conditions in the pollen-producing anther and germination-promoting conditions on the receiving stigma. Here we show that the differential environmental sensitivity of pollen germination in Arabidopsis thaliana is regulated by an AGC kinase cascade involving 3-PHOSPHOINOSITIDE-DEPENDENT KINASE 1 (PDK1) as upstream regulator, the AGC1.5 and AGC1.7 kinases as intermediate transducers and pollen germination promoting GUANINE NUCLEOTIDE EXCHANGE FACTORS (GEFs) as their likely phosphorylation targets. Loss of AGC kinase signaling resulted in precocious pollen germination in the anther, whereas AGC1.5/7 kinase overexpression or expression of a phosphomimic ROPGEF version inhibited pollen germination. Our data reveals that pollen germination requires the fine-tuned activity of an PDK1-AGC1 kinase cascade, preventing humidity-dependent precocious pollen germination on the anther.