Directed Evolution of Acoustic Reporter Genes Using High-Throughput Acoustic Screening.

A major challenge in the fields of biological imaging and synthetic biology is noninvasively visualizing the functions of natural and engineered cells inside opaque samples such as living animals. One promising technology that addresses this limitation is ultrasound (US), with its penetration depth of several cm and spatial resolution on the order of 100 µm. 1 Within the past decade, reporter genes for US have been introduced 2,3 and engineered 4,5 to link cellular functions to US signals via heterologous expression in commensal bacteria and mammalian cells. These acoustic reporter genes (ARGs) represent a novel class of genetically encoded US contrast agent, and are based on air-filled protein nanostructures called gas vesicles (GVs). 6 Just as the discovery of fluorescent proteins was followed by the improvement and diversification of their optical properties through directed evolution, here we describe the evolution of GVs as acoustic reporters. To accomplish this task, we establish high-throughput, semi-automated acoustic screening of ARGs in bacterial cultures and use it to screen mutant libraries for variants with increased nonlinear US scattering. Starting with scanning site saturation libraries for two homologs of the primary GV structural protein, GvpA/B, two rounds of evolution resulted in GV variants with 5- and 14-fold stronger acoustic signals than the parent proteins. We anticipate that this and similar approaches will help high-throughput protein engineering play as large a role in the development of acoustic biomolecules as it has for their fluorescent counterparts.

Genomically mined acoustic reporter genes for real-time in vivo monitoring of tumors and tumor-homing bacteria.

Ultrasound allows imaging at a much greater depth than optical methods, but existing genetically encoded acoustic reporters for in vivo cellular imaging have been limited by poor sensitivity, specificity and in vivo expression. Here we describe two acoustic reporter genes (ARGs)-one for use in bacteria and one for use in mammalian cells-identified through a phylogenetic screen of candidate gas vesicle gene clusters from diverse bacteria and archaea that provide stronger ultrasound contrast, produce non-linear signals distinguishable from background tissue and have stable long-term expression. Compared to their first-generation counterparts, these improved bacterial and mammalian ARGs produce 9-fold and 38-fold stronger non-linear contrast, respectively. Using these new ARGs, we non-invasively imaged in situ tumor colonization and gene expression in tumor-homing therapeutic bacteria, tracked the progression of tumor gene expression and growth in a mouse model of breast cancer, and performed gene-expression-guided needle biopsies of a genetically mosaic tumor, demonstrating non-invasive access to dynamic biological processes at centimeter depth.

Demonstration of a Longitudinal Medical Education Model (LMEM) Model to Teach Point-of-Care Ultrasound in Resource-Limited Settings.

Background: Short-term medical missions prevail as the most common form of international medical volunteerism, but they are ill-suited for medical education and training local providers in resource-limited settings. Objective: The purpose of this study is to evaluate the effectiveness of a longitudinal educational program in training clinicians how to perform point-of-care ultrasound (POCUS) in resource-limited clinics. Design: A retrospective study of a four-month POCUS training program was conducted with clinicians from a rural hospital in Haiti. The model included one-on-one, in-person POCUS teaching sessions by volunteer instructors from the United States and Europe. The Haitian trainees were assessed at the start of the program and at its conclusion by a direct objective structured clinical examination (OSCE), administered by the visiting instructors, with similar pre- and post- program ultrasound competency assessments. Results: Post-intervention, a significant improvement in POCUS competency was observed across six different fundamental areas of ultrasound (p < 0.0001). According to our objective structured clinical examination (OSCE), the mean assessment score increased from 0.47 to 1.68 out of a maximum score of 2 points, and each trainee showed significant overall improvement in POCUS competency independent of the initial competency pre-training (p < 0.005). There was a statistically significant improvement in POCUS application for five of the six medically relevant assessment categories tested. Conclusion: Our results provide a proof-of-concept for the longitudinal education-centered healthcare delivery framework in a resource-limited setting. Our longitudinal model provides local healthcare providers the skills to detect and diagnose significant pathologies, thereby reducing avoidable morbidity and mortality at little or no addition cost or risk to the patient. Furthermore, training local physicians obviates the need for frequent volunteering trips, saving costs in healthcare training and delivery.