Ultrasound is among the most widely used non-invasive imaging modalities in biomedicine, but plays a surprisingly small role in molecular imaging due to a lack of suitable molecular imaging agents. Although conventional microbubble contrast agents are gaining acceptance in non-invasive diagnosis of certain cardiovascular diseases and cancers, they have limited utility as labels of specific cells and tissues outside thebloodstream because their micron size typically confines them to the blood stream. As a result, ultrasound has yet to fulfill its full potential to enable convenient, rapid molecular imaging in biomedical research and potential clinical areas including cancer, immunology, neurology and infectious disease. We propose to address this need by borrowing from nature. Specifically, we will develop molecular imaging agents based on a unique class of genetically encoded gas nanostructures known as gas vesicles (GVs). Expressed by aquatic microorganisms as a means to control buoyancy, GVs are hollow protein-shelled compartments 50-500 nanometers in size that exclude water but are permeable to gas. Unlike artificial micro bubbles, GVs are not pressurized and allow gases to freely exchange with the surrounding medium. This results in a very stable nanoscale configuration enabling a broader range of potential molecular imaging applications. In preliminary results, we have demonstrated that GVs from multiple species produce stable ultrasound contrast that is readily detected in vitro, inside cells and in vivo. Thefact that GVs are genetically en- coded provides an unprecedented opportunity of engineering their properties at the genetic level to optimize their acoustics, biodistribution and targeting fo specific applications. In addition, there is the potential of adapting GVs as reporter genes - for the first time combining the ability of ultrasound to image at depth in vivo with the ability of genetic reporters to directly visualize cellular events such as gene expression. To address our hypothesis that GVs can serve as versatile molecular imaging reporters for ultrasound, we propose to develop this new class of molecular imaging agents by (1) understanding GVs' genetically en- coded acoustic properties through physical characterization and modeling, (2) using a genetic engineering plat- form to optimize GVs' acoustic, biological and targeting properties, (3) demonstrating the ability of these nanostructures to target and image extravascular tumor cells in vivo and (4) expressing GV-forming genes in mammalian cells. Successful completion of this project will result in a transformative advance in molecular imaging with ultrasound: a fundamentally new class of stable, nanosized, genetically tunable, molecularly targetable extra- vascular imaging agents, with immediate relevance in biomedical research and the potential for future clinical translation. In addition, this work will stimulate advances in biophysics, molecular and cellular engineering and imaging technology that will contribute more generally to biomedical imaging and bioengineering research.