Majors of undergraduate participants
Computer Science, Physical Sciences
- 13 Interdisciplinary Research teams comprised of faculty, graduate students and undergraduates, with half of each team from life sciences and half from other STEM disciplines
- Eight of the 25 faculty mentors were assistant professors
- Of 33 undergraduates, 6% of undergraduates were sophomores, 46% were juniors, and 48% were seniors
- 45% of the students were under-represented in their disciplines
- 30% were from underrepresented minority groups
- 15% were women with majors in engineering, geology, physics and chemistry
- 24 graduate students participated as research mentors to the undergraduates
- With funds from SUNY Research Foundation, three additional projects were added, supporting 6 additional undergraduates, bringing this years total to 39
1. Investigation of the pharmacokinetics of pHLIP conjugates using in vitro and in silico models | Read more
2. Integrating, metal, carbon and nitrogen biochemistry in two types of roadside ecosystems | Read more
3. Glutamate Transport and Nociception | Read more
4. Ancient Microorganisms in Fluid Inclusions in Halite and Gypsum | Read more
5. Lyme Disease Epidemiology | Read more
6. Image Processing to Characterize In Detail the Spatial Properties of Multiple Bacterial
within Biofilm | Read more
7. Statistical-Computer Based Technology in Developing a Clinical Decision Support System for Malignant Pleural Effusion Analysis | Read more
8. Characterization of Biofilm Formation on Sensors | Read more
9. Effects of Atrazine Exposure on Genome-wide Expression Levels in
Drosophila melanogaster Females | Read more
10. Biosynthesis of Pseudomonas Quinoline Signal and integration into fatty acid metabolism in Pseudomonas aeruginosa | Read more
11. Use of Engineered Nanovesicles to Investigate Formation and Biological Function of Bacterial Outer Membrane Vesicles | Read More
12. The Genetic Basis to Variation in Male Courtship Song in Drosophila melanogaster | Read more
13. A Microfluidic Fluorescent Cholesterol Biosensor | Read more
14. Three-dimensional Imaging and Volumetric Analysis the famous Gilboa Trees | Read more
15. Statistical Learning Theory Application : a Potential Technology to Help Improve
Colonoscopy Scheduling Guidelines | Read more
16. The Role of Altered Biochemical and Biophysical Conditions on Heart Valve Endothelial
to Mesechymal Transformation | Read more
A physiologically based pharmacokinetic (PBPK) mathematical model describes an organism as a set of interconnected compartments based on vasculature structure, and is designed to describe the time-dependent distribution of a chemical or drug in various tissues. This type of model is based on known physiological parameters such as blood flow rates, tissue volumes, and chemical characteristics such as partition coefficients and metabolic rates. A PBPK model can be useful for cross-species extrapolation and as a human surrogate for estimating risks associated with chemical or drug exposure. A cell culture analog (CCA) is a physical representation of a PBPK model. The CCA devices consist of channels and chambers arranged and sized to mimic the residence time and flow distribution of the corresponding PBPK model. Where the PBPK model mathematically specifies an organ or tissue compartment the CCA has an actual chamber holding a cell type that mimics the organ or tissue, while recirculating culture medium represents the circulatory system. The goal of a CCA is to create an in vitro system that can replicate some of the cell-cell interactions (i.e. interactions through soluble proteins and metabolites) in humans or animals not easily studied in vivo or in silico and to apply these observations to toxicology studies. Results obtained with CCA devices can also help to refine the corresponding PBPK models. One of the goals of this project is to develop a PBPK model and corresponding CCA microfluidic device that can describe the mechanism of action and toxicity of pHLIP and pHLIP-toxin conjugates.
Cancer chemotherapy is limited by toxic effects in healthy tissues, motivating the search for targeted drug delivery.We have developed an approach to target acidic solid tumors using the pH-(Low) Insertion Peptide (pHLIP) (Figure 1a). So far we have shown that pHLIP can translocate phalloidin, a membrane-impermeable polar toxin (Log P -2), into cancer cells to inhibit proliferation in vitro (Figure 1b)4. In collaboration with the Mahler lab, work proposed here aim to investigate the pharmacokinetics of pHLIP conjugates using physiologically realistic models based on microscale cell culture analogs (µCCA).
While pHLIP is a promising system, studies in mice revealed the following shortcomings that point to directions of further studies. I will discuss these directions in the context of µCCA collaboration:
(a) Phalloidin is probably not a sufficiently potent cytotoxic agent in vivo. For this reason, we will synthesize pHLIP-conjugates using cytotoxic agents far more potent than phalloidin, such as calicheamicin, auristatin, maytansinoid DM4, and CC-1065/adozelesin analog DC1. Unlike phalloidin (Log P ~ -2), these toxins are lipophilic (Log P ~ 1-5) and membrane-permeable. Attaching pHLIP to these toxins serve two purposes in vivo: (1) pHLIP translocation may circumvent capture of the drug by P-glycoproteins —the drug efflux pumps responsible for multi-drug resistance in cancer cells; and (2) at neutral pH, pHLIP may partition to the cell surface but will not insert, effectively blocking toxin entry into healthy cells (and reducing off-site toxicity). The µCCA model seems the perfect setting for studying the following questions: How does pHLIP attachment alter the metabolic fate of toxin molecules of various lipophilicity (Log P of -3 to 5)? How does the chemical nature of the linker (e.g. disulfide, sterically-hindered disulfide, ester, or azo group) impact the metabolic fate of the toxin cargo in pHLIP conjugates?
(b) Pharmacokinetics of pHLIP monomer is poor. More than 90% of pHLIP peptides are lost within 24 hours. The µCCA model will allow us to probe the metabolic fate of pHLIP and pHLIP-conjugates in more detail (by answering questions such as 'Are pHLIP peptides cleaved during circulation after i.v. injection?' or 'Are pHLIP peptides specifically metabolized by certain cell populations?').
(c) The pHLIP peptide targets acidity in vivo. Thus, acidic tissues other than tumors will also be targeted, such as the kidneys (as we observed), which may lead to toxic side effects in pHLIP-mediated drug delivery. We propose to synthesize pHLIP-PEG conjugates in the size range of ~ 5-10 nm in hydrodynamic diameter (M.W. 30-100 kD). If the µCCA system can accurately model renal filtration (the size of the renal pore is ~ 5 nm), then the pharmacokinetics of pHLIP-PEG conjugates may be investigated.
Faculty Mentors: Gretchen Mahler (Bioengineering) and Ming An (Chemistry)
Graduate Mentor: Courtney Sakolish (Bioengineering)Back to top
Roadside ecosystems (areas adjacent to roadways) are exposed to various traffic related pollution including elevated input of reactive forms of nitrogen (N: mainly NH3 and NOx), metals (e.g., Zn, Cu, Cd), and road salt (as winter deicing agents). The transformation and transfer mechanisms of reactive N and metals are governed by complex biogeochemical interactions, with the ultimate fates of both N and metals in urban environments being highly uncertain. Along with road salt, both N and metal also affect the biogeochemical cycling of carbon (C). The cycling of C is controlled by plant fixation of CO2 to organic C and the microbial respiration of organic C and the releases of CO2 to the atmosphere, which can play an important feedback role in global climate change. Locally, the amount and availability of soil organic C could also affect the metal mobility and consequently groundwater contamination.
For Project I, we have archived soils from a three-year manipulation experiment where plots in an open-field roadside ecosystem next to I-81 had received experimental additions of nitrate, sodium chloride, and water (control). The experiments were conducted by Prof Zhu's lab in the Department of Biological Sciences. Salt input significantly reduced the plant photosynthesis and soil microbial respiration rates. Now we plan to investigate the impact of biogeochemical alterations of C, N, and salt on the downward metal mobility. Soils collected from the experimental plots will be analyzed for major traffic-related metals in Prof. Graney's lab in the Department of Geological Sciences. For Project II, we have ongoing studies in roadside ditches where we have identified contamination of soils by metals (plus high salt content!). Here we plan to investigate whether such contaminated soils from ditches would have reduced C content, microbial activity, and N transformation.
Students will conduct metal analyses (Zn, Cu, Cd, and other selected metals) in the Geology laboratories. They will use microwave digestion to separate metals from the soil matrix and then analyze metal concentrations using the ICP-OES and ICP-MS. In the Biology laboratories, they will analyze N content (particularly nitrate) using the Lachat autoanalyzer, and quantify soil C content and microbial use of C using the GC (gas chromatography). HHMI students will also participate in field work including collecting soils from both the I-81 site and ditches.
Faculty Mentors: Joseph Graney (Geology) and Weixing Zhu (Biology)
Graduate Mentors: Miki Smith (Geology) and Stephanie Craig (Biology)Back to top
The HHMI research project represents the collaborative efforts of chemist/biophysicist Christof Grewer and behavioral pharmacologist Jilla Sabeti to unravel the intricate role and mechanisms of glutamate homeostasis in pain sensory and behavior regulation. Glutamate neurotransmission in the central nervous system plays a critical role in many fundamental brain functions, including learning and memory and pain/reward sensory regulation. Excess extracellular glutamate leads to neurotoxicity. As such, considerable cellular resources are committed toward preserving extracellular glutamate homeostasis, most notably thru highly regulated cellular reuptake processes via glial/neuronal glutamate transporter proteins. The proposed research uses an interdisciplinary approach to explore the fundamental problem of understanding dysregulation of glutamate transporter activity as a critical mechanism in pathological pain sensory and behavioral regulation. This work will involve generation of rats that have been made dependent on the abused drug alcohol and, consequently, show a compensatory up-regulation in glutamatergic activity, a reported neuronal hallmark of alcohol addiction. At the systems level of analysis, work in Dr. Sabeti's lab will involve behavioral pharmacology techniques to measure alterations in thermal nociceptive responding in the alcohol-dependent rat model. The role of glutamate transporter activity in altered nociception will be evaluated on a hot plate apparatus by screening drugs reported to be potent activators of glutamate transporter expression, including b-lactam antibiotic, as well as inhibitors that down-regulate glutamate transporter function. Students will gain skill in behavioral pharmacology methodology and essential pharmacokinetic and pharmacodynamic principles. At a chemical and biophysical level of analysis, work in Dr. Grewer's lab will involve synthesis and biophysical characterization of novel, hydrophobic glutamate-transporter-specific pharmacological tools that have the ability to cross the blood brain barrier, in contrast to the traditionally-used, hydrophilic and highly-charged glutamate transporter antagonists. Furthermore, the mechanism of antibiotic-induced glutamate transporter up-regulation will be investigated, which may be through a combination of transcriptional, translational, and direct-functional effects. These experiments will require the student to learn techniques in synthetic organic chemistry, as well as functional characterization of membrane proteins through heterologous expression followed by electrophysiological analysis. This combined interdisciplinary approach spanning molecular biophysics to whole animal experimentation will provide key insights into pharmacological targeting of glutamate transporter activity as potential treatment strategy for pathological pain sensory processing related to dysregulated glutamate homeostasis.
Faculty Mentors: Jilla Sabeti (Psychology) and Christof Grewer (Chemistry)
Graduate Mentors: Miguel Cabrea (Psychology) and Rose Tanui (Chemistry)Back to top
The proposed research explores the fundamental problem of long-term survival of microorganism communities and preservation of biomaterials in fluid inclusions in halite and gypsum. The goal of the proposed research is to obtain data on the distribution, survival, and diversity of microorganism communities and biomaterials that have been in the subsurface for periods of thousands to hundreds of millions of years. It will involve systematic examination of the state of survival and preservation of the suite of microorganisms, including prokaryotes (Bacteria and Archaea), eukaryotes (algae, fungi) as well as DNA and other biomolecules (i.e., chlorophyll, carotenoids, glycerol) in fluid inclusion ecosystems in halite and gypsum. The research plan will follow the successful interdisciplinary approach recently used for the study of halite from the subsurface of Death Valley, CA, but extended to older halite deposits, 105 to 108 Ma in age, and for the first time to gypsum. New emphasis will be placed on: (1) Molecular biological techniques involving amplification of fragments of DNA by the polymerase chain reaction (PCR), followed by cloning and sequencing, which will characterize the phylogenetic diversity of microorganisms in fluid inclusions in saline minerals. (2) Raman spectroscopy, which has the potential to quantitatively characterize the nature of biomolecules in fluid inclusions, in situ.
Faculty Mentors: Tim Lowenstein (Geology) and J. Koji Lum (Anthropology/Biology)
Graduate Mentors: Sarah Feiner (Geology) and Yue Zhang (Anthropology)Back to top
Emerging and re-emerging infectious diseases in developed and developing nations where chronic disease is on the rise is a hallmark of the 4th epidemiologic era. Diseases such as Dengue Fever, West Nile Virus and Malaria are re-emerging in areas where either they had previously been eradicated or did not exist at all. Over the last four decades new diseases such as HIV-AIDS, Ebola hemorrhagic fever and Lyme disease are the result of new human-environment interactions. Lyme disease emerged in the mid 1970's and has become the most common vector-borne disease in the United States today. The Binghamton University Lyme disease research group is investigating the emergence of this disease in upstate New York and other areas of the Northeastern U.S. The disease occurs in humans when an infected tick comes into contact with a human being and attaches itself for a blood meal. In the process of feeding, the tick transfers a gram negative bacteria called a spirochete. The Lyme disease spirochete is known generally as Borrelia burgdorferi.Lyme disease ecology and evolution is driven by geographic, ecologic, climatic, and human land use patterns as they relate to the Lyme disease vector, Ixodes scapularis(black legged tick or deer tick).
The Binghamton University (BU) Lyme disease research group is composed of faculty, staff, undergraduate and graduate students. We are currently using the Binghamton University campus as a natural experimental model to study the factors involved in Lyme disease transmission in a small geographically defined area. The project centers on five strategic areas (field ecology, laboratory analysis, field behavioral study, clinical symptomology and mathematical modeling) that are designed to address the factors influencing Lyme disease and other tick-borne infection emergence in human populations in upstate New York and other northeast areas. The field and laboratory component of the project involves the collection of ticks from various microecologies on campus using a corduroy cloth to drag the leaf litter and lower vegetation. Thus far we have collected and tested 411 ticks from the local region, including 310 on the Binghamton University campus of which 210 are larvae that have not yet taken a blood meal and thus not expected to harbor any mouse-borne infectious agents. Total DNA is extracted and the presence of B.burgdorferi, A. phagocytophilum, B. microti, and R. rickettsii is determined by PCR amplification. Demography, human traffic patterns, the built environment and human behaviors are also currently being assessed to determine exposure and risk that will lead to subsequent public health strategies and interventions.
Modeling of Lyme disease risk factors and the Lyme epidemic is a major research component of the project. Modeling of ecological dynamics of Lyme disease is an emerging subfield within the broader context of Lyme disease epidemiology. Models relating to vector-host-environment interactions are at the core of understanding the interrelating variables involved in disease transmission to humans. Clinical diagnosis of Lyme disease is heavily reliant on serologic and objective testing and doesn't necessarily incorporate ecological factors. Current Lyme disease laboratory diagnostic measures can be subject to error as well as untimely with a window of effectiveness which peaks four weeks after infection. We propose utilizing variables derived from existing models combined with epidemiological, ecological, social and behavioral factors to formulate a standardized algorithm of risk for use at the clinician level.
The methodological approach involves variables used in ecological models for predicting spatial distribution of disease-causing ticks. The variables include geographical location, annual precipitation, temperature, vegetation, land-use patterns and season. These ecological variables will be combined with a mathematical/computational dynamical model of spatial distribution of ticks as well as other epidemiological variables of age, sex, occupation and recreational habits. In addition, local incidence and engagement in risk prone activities will be integrated into the model. These variables will be weighted according to their predictive value in the models and combined to yield a composite score. The score will be normalized against a scale of 0 to 10 where 0 indicates the least risk and 10 indicates the maximum risk of infection. We expect that individuals with scores approaching 10 will be at the highest risk for being exposed to the pathogen that causes Lyme disease, Borrelia burgdoferi, while those scoring low will be at minimal risk.
Assigning risk scores to individual patients is important in determining further diagnostic or treatment protocols for a significant number of patients who do not present with classic Lyme disease symptoms and objective findings. Such a tool would add to the diagnostic armamentarium in Lyme disease detection and potentially speed the diagnosis and treatment of patients outside the classic presentation of Lyme disease. Investigative research will need to be conducted regarding the importance of each variable in the incidence of Lyme disease. This will allow for a more accurate scoring model when assessing a variable's weight of risk with respect to one another.
Faculty Mentors: Ralph Garruto (Anthropology) and Hiroki Sayama (Bioengineering)
Graduate Mentors: John Darcy (Anthropology) and Jeff Schmidt (Bioengineering)Back to top
The proposed research will investigate the dynamics of the multi species biofilm populations within Eukaryotic epithelial cell lines. Microbial infections, including nosocomial infections are mainly caused by multi-species bacterial populations. However, the available treatments are based on studies of single species bacterial populations. Currently not much is known about the initial stages of infection and how several bacterial species interact and eventually develop into biofilm within a Eukaryotic cell. Understanding the development of mixed species biofilms within epithelial cells could lead to a significant improvement on the efficacy of the treatment of infections.
This research will involve the tracking of individual cells to better understand the mechanisms by which they attach to a Eukaryotic cell and the interaction of the two bacterial species within the biofilm population and with the epithelial cells. Biofilms composed of Pseudomonas aeruginosa and Staphylococcus aureus will be used to co-infect Eukaryoticepithelial cells. We will monitor, by microscopy, the development of the biofilms throughout the infection of the Eukaryotic cells. Image processing, estimation and optimal hypothesis testing algorithms will be employed to accurately estimate population density and distribution of different bacterial species in the spatial domain and to develop a predictive model for the effect of local population density of one species versus another. This includes the continued development of a software product for analyzing bulk image sets acquired using confocal microscopy. Once the basic methods are established further research will be done to establish the efficacy of certain antimicrobials in biofilm killing and the change in shifts in population dynamics occurring throughout the treatments.
Faculty Mentors: Claudia Marques (Biology) and Scott Craver (Electrical and Computer Engineering)
Graduate Mentors: Alireza Farrokh-Baroughi (Electrical and Computer Engineering) and Aleksey Morozov (Biology)Back to top
The HHMI research project brings together Research Professor Walker Land , Dave Schaffer, Xingye Qiao, Dr. Sidhu, Chris Paquette and Martha Nelson who have been and are developing statistical learning theory (SLT) paradigms that result in complex adaptive systems (CAS), which measure the efficacy of drug treatments, survivability analysis of patients with operable metastatic lung cancer, and developing methods to intelligently process covariant features to accurately ascertain efficacy of new biomedical density measurements as applied to cancer screening and diagnosis. Specifically this project involves developing a Clinical Decision Support System (CDSS) for Malignant Pleural Effusion Analysis
The purpose of this hypothetical clinical decision support system is to aid clinicians in making the correct diagnosis in situations where malignancy may be a likely cause of patient symptoms/signs. While the precise clinical question to be answered is unknown this system has the potential to deliver highly accurate information pertaining to the differentiation between malignancy and other causes (diagnosis). The output information can be ideally used by the pathologist as a second opinion to check against their conclusions.
We will rely on our assisting pathologist to aid us in reviewing our system and logic to ensure that we fully understand the question/issue at hand, and more importantly, create a system that can deliver value to both patients and other clinicians. In my brief review, malignant and benign diagnosis seem to be the only situations under investigation in the academic realm. The assistant pathologist will provide us with critiques on what features we should consider in our system, what disease we should focus on, how the flow of our system will impact work-flow in clinical settings, and whether our system is scalable—just to name a few.
- Laboratory Pleural Effusion Analysis (Data Collection)
- Computational Pleural Effusion Analysis (Data Analysis)
- Diagnostic Output for Individual Patient (Classification)
Faculty Mentors: Walker Land (Bioengineering), David Schaffer (Bioengineering) Xingye Qiao (Mathematics)
Graduate Mentor: Yinglei Li (Bioengineering)Back to top
The HHMI research project brings together a chemist Wayne Jones and biologist, Karin Sauer, to explore capabilities of the bacterium Pseudomonas aeruginosato form surface associated communities called biofilms. The Jones group has designed new conjugated fluorescence turn-on polymers which capitalize on recent discoveries in inorganic/organic hybrid systems. The systematic characterization of the selectivity and sensitivity for these materials with known toxic metal ions in aqueous solution (Fe2+/3+, Hg+, Crn+, Pb+/2+, and Cd+/2+) will lay the ground work for next generation sensors for application to anions, toxic small molecules, or biomarkers of interest to the environmental health science community. Our primary objective is to develop the next generation of fluorescent polymeric organic/inorganic hybrid chemosensor materials focusing on turn-on fluorescent polymer chemosensors (FCP's) with improved selectivity for Fe2+/3+. Preliminary studies have demonstrated great success with one such "turn-on" organic/inorganic hybrid polymer chemosensor; tmeda-PPETE/Cu2+ was found to be highly sensitive (10 ppb) and selective for Fe2+ and Fe3+cations in hydrophilic solution. This will be applied to the study of biofilms and heme binding protein as described above. In addition, UV-vis, FTIR, NMR, and Electron Microscopy will be used as additional approaches to exploring the oxidation state of Fe in these systems. The functionality of these sensors in environmental and industrial settings depends on the surface of sensors remaining intact and not being subject to biofouling, a process involving the accumulation of bacteria (biofilm formation) on sensor surfaces. To determine which polymers, material composition and fiber thickness is the least subject to biofouling, biofilm formation on various sensors will be analyzed by viable count, microscopy including fluorescent and confocal microscopy, and various biochemical assays.
Faculty Mentors: Karin Sauer (Biology) and Wayne Jones (Chemistry)
Graduate Mentors: Yi Li (Biology) and Megan Fegley (Chemistry)Back to top
Atrazine is a commonly used herbicide yet is a potential endocrine disruptor that can have detrimental effects in non-target species. This HHMI research project brings together biologist Anthony Fiumera and statistician Xingye Qiao to identify genes which are differentially expressed in female D. melanogasterafter exposure to atrazine. Females from a standard laboratory strain of Drosophila will be reared on food containing 2ppm atrazine or control food. cDNA from six replicates of the control and exposed females will be sequenced using RNA-seq to estimate genome-wide expression levels. Statistical estimators will then be used to determine which genes are differentially expressed after exposure to this environmental toxicant. This would then be followed up with qRT-PCR, a method to carefully validate a subset of the important genes identified by the RNA-seq analysis.
Faculty Mentors: Anthony Fiumera (Biology) and Xingye Qiao (Mathematics)
Graduate Mentors:Sarah Marcus (Biology) and Yilin Zhu (Mathematics)
Bacterial cell-to-cell communication, termed Quorum Sensing (QS), is a system where bacteria secrete small molecule signals into the environment that are then detected by neighboring cells, which respond by altering gene expression and therefore behavior. Many of the gene expression and behavioral changes controlled by QS are found to be associated with bacterial virulence . In the opportunistic human pathogen Pseudomonas aeruginosa, the QS signal PQS not only induces the expression of virulence factors, but also physically promotes the packaging of these factors into Outer Membrane Vesicle (OMV) delivery vehicles for trafficking to target cells . This mode of transport allows pathogens to move disease-causing factors directly to target cells in a manner where they are protected from destruction by the immune system. Understanding the synthesis and function of PQS may therefore provide an exciting avenue toward developing therapies to mitigate P. aeruginosa virulence.
Genetic studies have identified genes involved in PQS biosynthesis and signal recognition . Based upon sequence analysis of these genes, a pathway has been proposed for the synthesis of PQS (Fig 1) . Anthranilate is believed to be condensed in a head-to-head fashion with -keto-decanoyl acid to form HHQ, which is then hydroxylated to form PQS. The route anthranilate takes through the pathway has been explored [3, 7, 8], but the reactions (and even the source) of the fatty acid substrates remain poorly understood. The question becomes even more interesting when one realizes that more genes exist in the biosynthetic gene cluster than are necessary for the synthesis of PQS. PqsD, PqsB and PqsC share sequence homology with one another and have been suggested to carry out condensation and fatty-acyl-carrier functions. However, as Fig 1 indicates, a maximum of two such proteins would be necessary to synthesize PQS. As few as one protein could be required for this since fatty-acyl-condensing enzymes often also function as carriers. This project proposes genetic experiments combining knockouts of pqsB, pqsC and pqsD in combination with disruptions in known sources of β-keto-fatty acids to identify the physiological source of the fatty acid substrate and identify whether the apparent redundancy in carriers reflects biological control over substrate flux from different sources. One argument in favor of this hypothesis is the observation that disruption of rhamnolipid synthesis (a potential source of fatty acid) results in partial loss of PQS production . An alternative explanation for the apparent protein redundancy is that the presence of multiple carrier proteins reflects a need for flexibility in the substrates used for PQS biosynthesis. Interestingly, it has been observed that P. aeruginosa naturally secretes two predominant forms of PQS, one with a 7-carbon alkyl substituent and one with a 9-carbon substituent . According to the proposed model for PQS biosynthesis, this would require the ability to use two different fatty acid substrates in the synthesis of PQS. Furthermore, feeding the organism fatty acid synthesis precursors caused it to shift production toward 9-carbon intermediates , suggesting that a control mechanism might exist to selectively produce different PQS analogs in response to the environment. Such control could be exercised through the use of specific carrier proteins for the substrates leading to the 7- and 9-carbon alkyl chain PQS molecules. To test this hypothesis, the proposed project will include the purification of PqsB and PqsC to investigate their specificity for carrying fatty acid substrates of different composition. Another aspect of this section of the project will be to investigate each of the three proteins for condensing activity. While PqsD has been shown convincingly to condense anthraniloyl-CoA with manlonate substrates (very short-chain fatty acids) to form a different secondary metabolite of P. aeruginosa , attempts to show condensation with the long-chain fatty acids required for PQS biosynthesis have resulted in minimal success (catalytic efficiency 5000-fold less than with malonyl-CoA) . Thus, either the true condensing activity for PQS biosynthesis resides in another protein, or previous PqsD studies have not been using the physiologically relevant fatty acid substrate. The proposed work will allow us to test both hypotheses: by looking for condensing activity in PqsB and PqsC as well as PqsD, and by using the PqsB- and PqsC-fatty-acid complexes (likely the true physiological substrates) as substrates for PqsD once their fatty acid specificities have been identified. It is clear that many important questions remain concerning the pathway toward PQS biosynthesis. The proposed project aims to tackle these questions through a combination of genetics and biochemistry. However, the biochemical analysis cannot proceed in the absence of collaboration with a skilled chemist to both synthesize substrates and assist in the characterization of the purified proteins.
To probe the substrate preference of PqsB, PqsC, and PqsD (with regard to carbon chain
length), we plan to synthesize various β-keto-acids that (in principle) would give
HHQ and PQS with 3, 5, 7 and 9 hydrocarbon side-chains. These enzyme-catalyzed reactions
will be monitored via UV-vis spectroscopy and HPLC (by following the formation of
HHQ). Ethyl 3-oxohexanoate is commercially available, which upon saponification would
provide the β-keto-acid corresponding to the C3 HHQ/PQS product. The syntheses of
3-oxo-octanoic acid (for C5 HHQ/PQS), 3-oxo-decanoic acid (for C7 HHQ/PQS), and 3-oxo-dodecanoic
acid (for C9 HHQ/PQS) will be carried out according to published procedures involving
the enolate of methyl acetate and the appropriate acyl chlorides, followed by hydrolysis
of the esters under basic conditions [1, 7]. To further investigate the mechanism
of HHQ formation, we also plan to synthesize CoA and N-acetyl cysteamine thioesters
of these β-keto-acids as mimics of the ACP-bound β-keto-acids . These synthetic
efforts would also allow us to radiolabel the β-keto-acid substrates if such material
is needed for future studies. In short, understanding the mechanism of HHQ/PQS formation
will lead to the design and synthesis of enzyme inhibitors as potential antimicrobial
agents. Further, elucidating the full biosynthetic pathway will then allow us to enzymatically
synthesize a multitude of PQS analogs for functional studies.
Faculty Mentors: Jeffery Schertzer (Biology) and Ming An (Chemistry)
Graduate Mentor: Joab Onyango (Chemistry)
Vesicle-mediated transport was once believed to be an exclusive feature of eukaryotic
systems. It is now understood that Gram-negative bacteria also have the ability to
sort, package and transport cargo to other cells in vesicles that bud off from their
outer membrane (OM). This mode of transport can provide target specificity, allow
for concentrated long-distance delivery and provide protection from the environment,
including elements of the immune system. As a testament to their versatility and importance,
OMVs have been shown to play important roles in the delivery of virulence factors,
modulation of the host immune system, bacterial cell-to-cell communication (termed
quorum sensing, QS), and the maturation of biofilms. For this reason, it is of great
interest to understand how OMVs are formed and the nature of their functions in bacterial
physiology and behavior.
In the opportunistic human pathogen Pseudomonas aeruginosa it was shown that the OMV-transported QS signal PQS (Pseudomonas Quinolone Signal) is responsible for inducing the formation of the very OMVs into which it is packaged , even in the absence of its receptor or de novo protein synthesis. This suggested that the OMV-inducing characteristic was based on a physical effect of PQS and not upon its role in altering gene expression as a QS signal. Following this, Dr. Schertzer proposed a mechanism (Fig 1) by which insertion of PQS into the outer leaflet of the outer membrane would expand that leaflet relative to the inner one and thereby induce membrane curvature, eventually leading to bleb formation and finally release of OMVs . This work successfully showed that PQS could induce curvature in model phospholipid membranes, but suffered from the caveat that it could not be tested using artificial membranes matching the asymmetric lipopolysaccharide (LPS)/phospholipid (PL) (outer leaflet/inner leaflet) architecture of the bacterial OM. This was an unfortunate limitation of technology since no such membrane structures have been made at the appropriate size or material scale to test these questions. Dr. Chiarot's nanoscale vesicle technology (below) promises to alleviate this problem, allowing OMVs of controlled size to be fabricated with lipid content and architecture identical to that of the natural OM. The proposed work will use large (bacteria-sized) engineered OMVs to test the applicability of the bilayer-couple model to bacterial surfaces. Nanoscale (50 – 500 nm) OMVs will be used to investigate the preference of PQS to insert into membranes of different curvature (as in ). That PQS might prefer curved membranes is a hypothesis developed in our group to explain the apparent ability of PQS to self-aggregate in the OM to be packaged into OMVs – greater than 80% of PQS in a bacterial culture is associated with OMVs. In addition, by engineering the OMVs to have the opposite asymmetry (LPS = inner, PL = outer), we will use engineered OMVs to screen for compounds that can expand the PL leaflet, which would antagonize natural OMV production and hinder bacterial virulence.
The proposed work also aims to define the role of OMVs in biofilm maturation. Biofilms are communities of bacteria that have attached to a surface and encased themselves in a structural and protective extracellular matrix. In this mode of growth, bacteria more resemble a multicellular tissue than the solitary planktonic cells that they are commonly thought to be. In fact, it is now appreciated that biofilm growth is the common mode of growth for bacteria in nature and biofilm infections make up the majority of bacterial infections. When PQS production is disrupted in P. aeruginosa, the organism fails to form mature biofilms , but it is unclear whether this is due to the loss of the QS function of PQS or the potential structural function of OMVs in the biofilm. Since PQS physically stimulates the formation of OMVs in P. aeruginosa, it is exceedingly difficult to generate natural OMVs that lack PQS to attempt to discriminate between these two functions. In addition to investigating PQS-free OMVs from other species, we will use engineered OMVs to attempt to complement biofilm maturation deficiencies in PQS biosynthetic mutant strains. These OMVs will contain no PQS or bacterial protein and will serve as the perfect 'blank slate' to add back individual components and identify key contributors to biofilm maturation.
The use of microfluidic technology is an attractive option for producing customized
synthetic vesicles. This technology has many important advantages, including: precise
control over dispensed volumes, high-throughputs, and repeatability. The internal
volume of a microfluidic device is on the order of nanoliters, while typical solution
flow rates can be as low as picoliters-per-minute. Smaller volumes mean less material
(i.e. lipid) is consumed – a significant issue for high-cost lipids such as LPS and
a true hurdle overcome using this technique. Most notably, our proposed technique
will be capable of building vesicles with asymmetric membranes, where each leaflet
is composed of a different lipid. This is an essential requirement for each of the
Our proposed technique uses liquid emulsions and lipid self-assembly inside a microchannel network built using a layer of poly-dimethylsiloxane (PDMS) bonded to a glass substrate with patterned electrodes . This system is ideal for achieving control over vesicle unilamellarity, size, and uniformity. Membrane curvature (proportional to vesicle size) will be tightly controlled, with both narrow and wide OMV size distributions possible. This feature is an excellent fit for downstream applications as both wide distributions (for studying PQS insertion into OMVs of different curvature) and narrow distributions (for biofilm complementation experiments) will be called for. Our strategy for forming vesicles relies on four key steps as shown in Fig 2: (i) emulsion formation using flow focusing, (ii) formation of co-flowing laminar streams (iii) lipid monolayer self-assembly at multiple liquid-liquid interfaces, and (iv) dielectrophoretic steering to transfer lipids to the emulsion surface. Lipid bilayer self-assembly takes place at the surface of the emulsions, which act as a template for the vesicle membrane. The emulsion itself forms the body of the vesicle, while steering of the emulsion using dielectrophoretic force allows the outer leaflet to be "painted" onto the inner leaflet as the emulsion slips through the interface between oil and aqueous phases. Typical emulsion (vesicle) diameters are ~10μm; however, diameters <1μm are achievable with the addition of an electric field to break up larger emulsions. For additional flexibility, the device can be fabricated to allow for the addition of exogenous materials (ie PQS, OMV proteins) prior or subsequent to their formation and also allow for on-chip analysis of the OMVs as they exit.
The proposed microfluidic device provides a flexible, reliable and low cost way to produce vesicles to biological specifications at high throughput. These tailored OMVs will then be used to investigate exciting biological questions that were technically unfeasible even in the recent past.
Faculty Mentors: Jeffery Schertzer (Biology) and Paul Chiarot (Mechanical Engineering)
Graduate Mentors: Alex Nello (Biology) and Li Lu (Mechanical Engineering)
Understanding the genetic basis to complex behaviors, such as courtship, can help us understand neural development and gene regulatory networks. Using genetic variation present in natural populations to study courtship can also allow us to study how natural selection might affect behavior. This HHMI project brings together Dr. R. Miles, Dr. C. Miles and Dr. A. Fiumera to study the genetic basis to variation in male courtship song in Drosophila melanogaster. This project will survey courtship song from more than 150 genetically distinct D. melanogaster lines that have their full genomes sequenced and genome wide association mapping will be used to identify which genes are affecting song parameters such as frequency or pulse rate.
Faculty Mentors: Anthony Fiumera (Biology), Carol Miles (Biology) and Ron Miles (Mechanical Engineering)
Graduate Mentor: Tim Galati (Mechanical Engineering)Back to top
Cholesterol garners attention for its role in cell membrane structure and steroid hormone biosynthesis, while also serving as a biomarker for lipid disorders such as atherosclerosis. As a consequence, methods to detect and monitor cholesterol have application to basic and to clinical sciences. Cholesterol detection has traditionally relied on coupling a multistep enzymatic cascade to a quantifiable electrical signal (Arya et al. 2008). Despite two decades of development, the enzyme-cascade approach to cholesterol biosensing remains burdened by interference from redox-active metabolites and enzyme instability.
This project seeks to develop a simple fluorescence based cholesterol biosensor. In place of the enzyme cascade, we will use a single autocatalytic protein that separates into two fragments in the presence of cholesterol. Cholesterol serves as a nucleophile, attacking at a precise location (Gly-Cys) in the protein sequence (Figure 1, left) (Porter et al. 1996). To repurpose this autocatalytic event into a cholesterol biosensor, we will modify the protein with fluorescent probes and immobilization tags so that self-cleavage can be monitored optically in real time. A prototype of the sensing platform, which we will be constructed using nanofabrication technology, is depicted below (Figure 1, right).
Faculty Mentors: Stephen Levy (Physics) and Brian Callahan (Chemistry)
Graduate Mentors: Steven Button (Physics and Timothy Owen (Chemistry)Back to top
Ongoing paleontological research from the Devonian Period (approx 385 million years before present) in the Catskill region of New York State has revealed critical new evidence about the world-famous "Earth's earliest forest" from the town of Gilboa in the Catskills of New York. The forest consists of sandstone casts of bases of Eospermatopteris trees standing in place. In 2007, we published a now widely circulated reconstruction of Eospermatopteris based on specimens collected nearby showing one or more trees lying on their sides. In 2012, we reported an astonishing discovery at the original site itself that the ancient forest floor was still intact. We mapped the site to determine the distribution of nearly 200 trees observed in this ancient forest. The next step in our research is to determine how these trees grew. This is important because although large, Eospermatopteris plants were not woody as might normally be expected of trees. Thus, understanding the basic physics of support, and mode of development to large size, becomes central to interpreting their ecological role in the ancient forest. To do this, we propose to image multiple stumps of different sizes, and often different shapes, to construct fully three-dimensional models of each. From this we plan to obtain volumetric information from base to top of each specimen to compare with metric information preserved about the plant's supporting network and vascular system. Interestingly, some trees appear to have been deformed either in life or during the preservation process, potentially providing important clues about the tree's structural integrity and mode of growth.
Students will be involved in collecting photographic data, consisting of 30-50 or more digital images from different points of view around each specimen. From the photographs, existing or newly developed software will be employed to construct fully three-dimensional virtual models. Select virtual models will then be used to make scale models of important specimens via digital 3D printing either in-house or using an online fabrication service. Volumetric data including total volume and volume by horizontal increment will also be collected from the virtual models and input into a spreadsheet along with other observational information. From this, statistical analysis will be performed to reconstruct trajectory of size/growth from base to tip for multiple specimens. Results of this work will represent a first of this kind in paleobotany.
Faculty Mentor: Lijun Yin (computer science)
Introduction: Current guidelines for colonoscopy intervals have been extrapolated using clinical experience, biological models for progression of adenomatous lesions, and clinical studies. With over a decade of experience performing screening colonoscopies, it is time to exploit these data to see if we can improve colonoscopy scheduling policies. Recent advances in artificial intelligence have produced powerful new algorithms for pattern discovery. Older versions of these algorithms have proven useful in bioinformatics.
Aim: To use statistical learning theory algorithms to assess efficacy of current colonoscopy intervals.
Methods: The medical records of patients receiving an initial screening colonoscopy between 1988 and 2012, and having at least one follow-up colonoscopy, were collected. We developed an algorithm to compute a patient's risk category (avg. above avg. high) using all information available after a colonoscopy. A data set was assembled with a record for each visit-pair containing all patient information from visit-1, the time interval between the visits, and the risk assessment computed at visit-2. Pattern discovery methods (bagging with decision tree learning and PLS) were applied to the task of predicting the future risk (visit-2) risk using information known at visit-1.
Year 1 Results: To date 491 such patients have been identified, predominantly Caucasian (84%). The ages range from 28 to 85 with a mean of 59.6, 48% female. The follow-up intervals range from two to 208 months, with a mean of 52.8. This yielded 875 visit-pairs, which were examined for the three cohorts corresponding to avg. above avg. and high risk patients at visit 1. The distributions of between-visit intervals were surprisingly similar for these cohorts. The strongest predictive pattern was found for the high risk cohort where cross-validation AUC was used for distinguishing those who would still be high risk at visit-2 from this who dropped to above avg. risk. The features contributing most to this prediction were: personal medical history, VCODE_13, Family history, height, weight AND BMI (i.e. all 3), number and types of polyps found, and 3 local neighborhoods. A modest, but rather lower predictability was found for the above avg. risk cohort and virtually no predictability for the avg. risk cohort.
Conclusion: This pilot study should provide the experience needed to design a larger study and estimate the sample sizes needed for statistical confidence. The possibility for evidence-based personalized colonoscopy scheduling is the dream.
Proposed Year 2 Activities: In the project's second year, two objectives will be pursued: samples that are more patient will be collected and pattern discovery activities will intensify.
1. The current set of just under 500 samples is inadequate to produce convincing findings, particularly when we divide patients into risk cohorts. Fortunately, the activities in the first year have yielded a how-to manual where the data collecting experiences have been captured. The next students only need to follow the procedures. The goal is to have 1000 by the end of year2.
(2.) Specific data mining activities will include (but not necessarily be limited to) validating that the initial results are still solid, examine the information on inter-visit interval, seek corroboration of the zip code findings (data have just be acquired from a local study of C incidents), and explore the use of PNN technology currently under development here at Binghamton University.
Specifically, a new PNN architecture ,which implements Bayesian processing with losses explicitly included (and the effect measured by external ROC analysis) as well as a new design, which produces 3 outputs taken 2 at a time simultaneously, has been developed and will be implemented using PYTHON. Also, a more conventional decision tree analysis is under consideration for implementation.
Faculty Mentors: Walker Land (Bioengineering) and J.David Schaffer (Bioengineering)
The goal of our research is to determine how and why calcific aortic valve disease is initiated. The aortic valve is exposed to a demanding physical and chemical environment and, under normal conditions, is continuously repaired by resident cells. One potential method of valve repair and an early event in valve disease is endothelial to mesenchymal transformation (EndMT), which was previously thought to occur only during embryonic valve development. We have developed cell culture models that mimic the physical and chemical environment of the aortic valve, and we study EndMT and how it relates to valve calcification. There are currently no options for treating calcific aortic valve disease except for valve replacement, and this research could lead to non-surgical therapies.
We have started to characterize some of the changes to the valve's physical and chemical environment that stimulate EndMT. Changes in the pattern of blood flow and high concentrations of inflammatory cytokines, for example, lead to an increase in this cell behavior. In this project we are focusing on how changes to the extracellular matrix that compose the aortic valve initiate mesenchymal transformation and how inflammatory cytokines and blood flow patterns then promote or deter EndMT and eventual valve calcification. We have developed tools including 3D cell culture scaffolds, microfluidic devices, and computational models to answer these questions.
Faculty Mentor: Gretchen Mahler (Bioengineering)
Graduate Mentor: Sudip Dahal (Bioengineering)