Samples were collected on cruise AE1409. Trichodesmium colonies were obtained by net tow (130 micron mesh) and serially washed in sterile surface seawater. Clean colonies were then incubated with or without a cocktail of quorum sensing molecules. After four hours of incubation, colonies were placed onto filters and stored in liquid nitrogen until RNA was extracted and submitted for sequencing at the Columbia University Genome Center. 

Methods: We extracted prokaryotic RNA from triplicate control and +AHL samples by first adding approximately 500 μL of glass beads to each cryotube and bead beating with a vortex adaptor for 5 minutes. We extracted total RNA using the yeast protocol from the Qiagen RNeasy Mini Kit with the added on-column DNase digestion using the RNase-free DNase Kit (Qiagen, Hilden, Germany). We processed DNase-treated total RNA through a MICROBEnrich kit following the manufacturer’s instructions (ThermoFisher Scientific, Waltham, MA, USA). We removed ribosomal RNA using a Ribo-Zero Magnetic Kit optimized for bacteria (Illumina, San Diego, CA, USA) following the manufacturer’s instructions. Finally, we purified the prokaryotic RNA extract using the RNeasy MinElute Cleanup Kit by following manufacturer instructions and eluting in 14 μL water (Qiagen). We pooled together triplicate samples, and pooled RNA extracts were quantified using the Take3 Nucleic Acid Quantification program and a Biotek plate reader. To further assess quality of pooled triplicate RNA samples, we used a BioAnalyzer and the RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, CA, USA). The JP Sulzberger Genome Center at Columbia University carried out RNA sequencing with a depth of 60 million paired end reads using an Illumina HiSeq protocol.

Quality control: We trimmed sequence reads and normalized following the Eel Pond Protocol for mRNAseq assembly. To obtain read counts for each sample, we mapped cleaned forward and reverse reads to metagenome assemblies from the same sampling locations that were previously characterized and clustered into orthologous groups (OGs). We carried out mapping using RSEM with the paired-end and Bowtie2 parameters. We summed counts for previously determined OGs for Trichodesmium and epibiont genome bins separately. We determined significant changes in OG relative abundance between control and +AHL samples by comparing control and sample treatments using a stringent empirical Bayes approach called Analysis of Sequence Counts (ASC). This approach evaluates the posterior probability associated with a given fold change across pooled triplicates, and performs similarly, but conservatively, on replicated and unreplicated sample datasets.  OGs were considered significantly higher or lower if they had a 95% or higher posterior probability of a fold change greater than 2 between treatment and control. Taxonomic relative abundance estimates for metagenome samples were previously calculated by multiplying the length of each contig in a genome bin by read mapping coverage, and then summing those values for all contigs. Please refer to the manuscript related to this metadata for more details and references.