Explore the vast range of titles available.

Research in science, technology, engineering, and mathematics (STEM) fields has become much more complex in the twenty-first century. As a result, the students of our Graduate School, who are all Ph.D. candidates, need to be trained in essential skills and processes that are crucial for success in academia and beyond. Some research problems are inherently complex in that they raise deep moral dilemmas, such as antimicrobial resistance, sustainability, dual-use research of concern (defined as well-intentioned scientific research that may be misused for nefarious purposes), and human cloning. Dealing with moral dilemmas is one of several core competencies that twenty-first-century Ph.D. students must acquire. However, this might prove difficult for STEM Ph.D. students who have had limited exposure to moral philosophy. Since the task of dealing with moral dilemmas in STEM research requires input from both scientific and philosophical disciplines, it is argued with the help of the 4 examples above that this task be explicitly modelled as an interdisciplinary process. Furthermore, it is argued that a particular model from the interdisciplinary education literature could serve as a learning tool to support ethical decision-making in research ethics and integrity courses for doctoral students.

Enterococcus faecalis is a Gram-positive, opportunistic, pathogenic bacterium that causes a significant number of antibiotic-resistant infections in hospitalized patients. The development of antibiotic resistance in hospital-associated pathogens is a formidable public health threat. In E. faecalis and other Gram-positive pathogens, correlations exist between lipid composition and antibiotic resistance. Resistance to the last-resort antibiotic daptomycin is accompanied by a decrease in phosphatidylglycerol (PG) levels, whereas multiple peptide resistance factor (MprF) converts anionic PG into cationic lysyl-PG via a trans-esterification reaction, providing resistance to cationic antimicrobial peptides. Unlike previous studies that relied on thin layer chromatography and spectrophotometry, we have performed liquid chromatography-tandem mass spectrometry (LC-MS/MS) directly on lipids extracted from E. faecalis, and quantified the phospholipids through multiple reaction monitoring (MRM). In the daptomycin-sensitive E. faecalis strain OG1RF, we have identified 17 PGs, 8 lysyl-PGs (LPGs), 23 cardiolipins (CL), 3 glycerophospho-diglucosyl-diacylglycerols (GPDGDAG), 5 diglucosyl-diacylglycerols (DGDAG), 3 diacylglycerols (DAGs), and 4 triacylglycerols (TAGs). We have quantified PG and shown that PG levels vary during growth of E. faecalis in vitro. We also show that two daptomycin-resistant (DapR) strains of E. faecalis have substantially lower levels of PG and LPG levels. Since LPG levels in these strains are lower, daptomycin resistance is likely due to the reduction in PG. This lipidome map is the first comprehensive analysis of membrane phospholipids and glycolipids in the important human pathogen E. faecalis, for which antimicrobial resistance and altered lipid homeostasis have been intimately linked.

The cell membrane plays a pivotal role in protecting bacteria against external threats, such as antibiotics. Cationic phospholipids such as lysyl-phosphatidyglycerol (L-PG) resist the action of cationic antimicrobial peptides through electrostatic repulsion. The bacterial cell membrane is an interface for cell envelope synthesis, protein secretion, virulence factor assembly, and a target for host cationic antimicrobial peptides (CAMPs). To resist CAMP killing, several Gram-positive pathogens encode the multiple peptide resistance factor (MprF) enzyme that covalently attaches cationic amino acids to anionic phospholipids in the cell membrane. While E. faecalis encodes two mprF paralogs, MprF2 plays a dominant role in conferring resistance to killing by the CAMP human β-defensin 2 (hBD-2) in E. faecalis strain OG1RF. The goal of the current study is to understand the broader lipidomic and functional roles of E. faecalis mprF. We analyzed the lipid profiles of parental wild-type and mprF mutant strains and show that while Δ mprF2 and Δ mprF1 Δ mprF2 mutants completely lacked cationic lysyl-phosphatidylglycerol (L-PG), the Δ mprF1 mutant synthesized ~70% of L-PG compared to the parent. Unexpectedly, we also observed a significant reduction of PG in Δ mprF2 and Δ mprF1 Δ mprF2 . In the mprF mutants, particularly Δ mprF1 Δ mprF2 , the decrease in L-PG and phosphatidylglycerol (PG) is compensated by an increase in a phosphorus-containing lipid, glycerophospho-diglucosyl-diacylglycerol (GPDGDAG), and D-ala-GPDGDAG. These changes were accompanied by a downregulation of de novo fatty acid biosynthesis and an accumulation of long-chain acyl-acyl carrier proteins (long-chain acyl-ACPs), suggesting that the suppression of fatty acid biosynthesis was mediated by the transcriptional repressor FabT. Growth in chemically defined media lacking fatty acids revealed severe growth defects in the Δ mprF1 Δ mprF2 mutant strain, but not the single mutants, which was partially rescued through supplementation with palmitic and stearic acids. Changes in lipid homeostasis correlated with lower membrane fluidity, impaired protein secretion, and increased biofilm formation in both Δ mprF2 and Δ mprF1 Δ mprF2 , compared to the wild type and Δ mprF1 . Collectively, our findings reveal a previously unappreciated role for mprF in global lipid regulation and cellular physiology, which could facilitate the development of novel therapeutics targeting MprF. IMPORTANCE The cell membrane plays a pivotal role in protecting bacteria against external threats, such as antibiotics. Cationic phospholipids such as lysyl-phosphatidyglycerol (L-PG) resist the action of cationic antimicrobial peptides through electrostatic repulsion. Here we demonstrate that L-PG depletion has several unexpected consequences in Enterococcus faecalis , including a reduction of phosphatidylglycerol (PG), enrichment of a phosphorus-containing lipid, reduced fatty acid synthesis accompanied by an accumulation of long-chain acyl-acyl carrier proteins (long chain acyl-ACPs), lower membrane fluidity, and impaired secretion. These changes are not deleterious to the organism as long as exogenous fatty acids are available for uptake from the culture medium. Our findings suggest an adaptive mechanism involving compensatory changes across the entire lipidome upon removal of a single phospholipid modification. Such adaptations must be considered when devising antimicrobial strategies that target membrane lipids.

Diabetic retinopathy (DR) is the foremost cause of blindness in people with diabetes worldwide, and early diagnosis is essential for effective treatment. Unfortunately, the present DR screening method requires the skill of ophthalmologists and is time-consuming. In this study, we present an automated system for DR severity classification employing the fine-tuned Compact Convolutional Transformer (CCT) model to overcome these issues. We assembled five datasets to generate a more extensive dataset containing 53,185 raw images. Various image pre-processing techniques and 12 types of augmentation procedures were applied to improve image quality and create a massive dataset. A new DR-CCTNet model is proposed. It is a modification of the original CCT model to address training time concerns and work with a large amount of data. Our proposed model delivers excellent accuracy even with low-pixel images and still has strong performance with fewer images, indicating that the model is robust. We compare our model’s performance with transfer learning models such as VGG19, VGG16, MobileNetV2, and ResNet50. The test accuracy of the VGG19, ResNet50, VGG16, and MobileNetV2 were, respectively, 72.88%, 76.67%, 73.22%, and 71.98%. Our proposed DR-CCTNet model to classify DR outperformed all of these with a 90.17% test accuracy. This approach provides a novel and efficient method for the detection of DR, which may lower the burden on ophthalmologists and expedite treatment for patients.

Macromolecular crowding (MMC) is a biophysical effect that governs biochemical processes inside and outside of cells. Since standard cell culture media lack this effect, the physiological performance of differentiated and progenitor cells, including extracellular matrix (ECM) deposition, is impaired in vitro . To bring back physiological crowdedness to in vitro systems, we have previously introduced carbohydrate-based macromolecules to culture media and have achieved marked improvements with mixed MMC in terms of ECM deposition and differentiation of mesenchymal stem cells (MSCs). We show here that although this system is successful, it is limited, due to viscosity, to only 33% of the fractional volume occupancy (FVO) of full serum, which we calculated to have an FVO of approximately 54% v/v. We show here that full-serum FVO can be achieved using polyvinylpyrrolidone (PVP) 360 kDa. Under these conditions, ECM deposition in human fibroblasts and MSCs is on par, if not stronger than, with original MMC protocols using carbohydrates, but with a viscosity that is not significantly changed. In addition, we have found that the proliferation rate for bone marrow-derived MSCs and fibroblasts increases slightly in the presence of PVP360, similar to that observed with carbohydrate-based crowders. A palette of MMC compounds is now emerging that enables us to tune the crowdedness of culture media seamlessly from interstitial fluid (9% FVO), in which the majority of tissue cells might be based, to serum environments mimicking intravascular conditions. Despite identical FVO's, individual crowder size effects play a role and different cell types appear to have preferences in terms of FVO and the crowder that this is achieved with. However, in the quest of crowders that we have predicted to have a smoother regulatory approval path, PVP is a highly interesting compound, as it has been widely used in the medical and food industries and shows a novel promising use in cell culture and tissue engineering.

Antimicrobial peptides (AMPs) are utilized by both eukaryotic and prokaryotic organisms. AMPs such as the human beta defensins, human neutrophil peptides, human cathelicidin, and many bacterial bacteriocins are cationic and capable of binding to anionic regions of the bacterial surface. Cationic AMPs (CAMPs) target anionic lipids [e.g., phosphatidylglycerol (PG) and cardiolipins (CL)] in the cell membrane and anionic components [e.g., lipopolysaccharide (LPS) and lipoteichoic acid (LTA)] of the cell envelope. Bacteria have evolved mechanisms to modify these same targets in order to resist CAMP killing, e.g., lysinylation of PG to yield cationic lysyl-PG and alanylation of LTA. Since CAMPs offer a promising therapeutic alternative to conventional antibiotics, which are becoming less effective due to rapidly emerging antibiotic resistance, there is a strong need to improve our understanding about the AMP mechanism of action. Recent literature suggests that AMPs often interact with the bacterial cell envelope at discrete foci. Here we review recent AMP literature, with an emphasis on focal interactions with bacteria, including (1) CAMP disruption mechanisms, (2) delocalization of membrane proteins and lipids by CAMPs, and (3) CAMP sensing systems and resistance mechanisms. We conclude with new approaches for studying the bacterial membrane, e.g., lipidomics, high resolution imaging, and non-detergent-based membrane domain extraction.

ABSTRACT Membrane vesicles (MVs) contribute to various biological processes in bacteria, including virulence factor delivery, antimicrobial resistance, host immune evasion and cross-species communication. MVs are frequently released from the surface of both Gram-negative and Gram-positive bacteria during growth. In some Gram-positive bacteria, genes affecting MV biogenesis have been identified, but the mechanism of MV formation is unknown. In Enterococcus faecalis, a causative agent of life-threatening bacteraemia and endocarditis, neither mechanisms of MV formation nor their role in virulence has been examined. Since MVs of many bacterial species are implicated in host–pathogen interactions, biofilm formation, horizontal gene transfer, and virulence factor secretion in other species, we sought to identify, describe and functionally characterize MVs from E. faecalis. Here, we show that E. faecalis releases MVs that possess unique lipid and protein profiles, distinct from the intact cell membrane and are enriched in lipoproteins. MVs of E. faecalis are specifically enriched in unsaturated lipids that might provide membrane flexibility to enable MV formation, providing the first insights into the mechanism of MV formation in this Gram-positive organism. Enterococcal MVs possess unique proteome and lipidome and activate NF-kB signaling in macrophages.

LLMs can provide substantial zero-shot performance on diverse tasks using a simple task prompt, eliminating the need for training or fine-tuning. However, when applying these models to sensitive tasks, it is crucial to thoroughly assess their robustness against adversarial inputs. In this work, we introduce Static Deceptor (StaDec) and Dynamic Deceptor (DyDec), two innovative attack frameworks designed to systematically generate dynamic and adaptive adversarial examples by leveraging the understanding of the LLMs. We produce subtle and natural-looking adversarial inputs that preserve semantic similarity to the original text while effectively deceiving the target LLM. By utilizing an automated, LLM-driven pipeline, we eliminate the dependence on external heuristics. Our attacks evolve with the advancements in LLMs and demonstrate strong transferability across models unknown to the attacker. Overall, this work provides a systematic approach for the self-assessment of an LLM's robustness. We release our code and data at https://github.com/Shukti042/AdversarialExample.

Vision-Language Models (VLMs) are now a core part of modern AI. Recent work proposed several visual jailbreak attacks using single/ holistic images. However, contemporary VLMs demonstrate strong robustness against such attacks due to extensive safety alignment through preference optimization (e.g., RLHF). In this work, we identify a new vulnerability: while VLM pretraining and instruction tuning generalize well to split-image inputs, safety alignment is typically performed only on holistic images and does not account for harmful semantics distributed across multiple image fragments. Consequently, VLMs often fail to detect and refuse harmful split-image inputs, where unsafe cues emerge only after combining images. We introduce novel split-image visual jailbreak attacks (SIVA) that exploit this misalignment. Unlike prior optimization-based attacks, which exhibit poor black-box transferability due to architectural and prior mismatches across models, our attacks evolve in progressive phases from naive splitting to an adaptive white-box attack, culminating in a black-box transfer attack. Our strongest strategy leverages a novel adversarial knowledge distillation (Adv-KD) algorithm to substantially improve cross-model transferability. Evaluations on three state-of-the-art modern VLMs and three jailbreak datasets demonstrate that our strongest attack achieves up to 60% higher transfer success than existing baselines. Lastly, we propose efficient ways to address this critical vulnerability in the current VLM safety alignment.