Life Sciences

Sam Passey, Lyndsey Graig, Christine Fiscer, RonJai Staton, Jeremy Scritchfield, Barbara Balbas, Amy Harmon, Craig Demke, Joey Jergins, Tim Bywater, and Dannelle Larsen Rife, Dixie State University Life Sciences Research in human development suggests relationships are vital for physiological and emotional well-being across the lifespan. Attachment theory is foundational for relationships and is intrinsic in human nature as it is represented through words of novelists. Attachments are developed within the first year of life based on caregivers’ appropriate, contingent, and prompt responses to the infant’s cues. Avoidant attachment develops when the infant receives minimal responses to his or her cues. John Bowlby proposed the attachment relationship between the infant and parent creates an internal working model (IWM). This IWM sets the foundation of all subsequent close relationships throughout the lifespan. Individuals who have avoidant attachment representations are dismissive of, and lack security in relationships. Living in a time where women were marginalized, segregated, and many lacked formal education, Mary Shelley effectively produced a popular work of fiction in the early 1800s. Shelley was a keen observer of relationships long before Attachment Theory was developed in the 1960s. Psychobiographical methods were used to examine Shelley’s Frankenstein as a case study of Attachment Theory. Results suggest Shelley’s Frankenstein depicts basic components of attachment theory, and “Frankenstein,” the monster character, exemplifies avoidant attachment. Through his dismissive and proximity seeking behaviors, the monster characterizes Bowlby’s description of avoidant attachment. Lacking relationships during critical periods for development of empathy, the monster loses the ability to feel remorse. This critical examination of early British literature as a case study for Attachment Theory lends retrospective support for the understanding of human relationships.

Avani Latchireddi, Wade Mather, and Joseph Tonna, University of Utah Life Sciences Purpose This research project is to assess patient satisfaction and feedback around care provided in the Surgical Intensive Care Unit (SICU) at the University of Utah Hospital with the goal of iterative improvement of care. Research Question/Hypothesis We hypothesize that by assessing patient satisfaction of ICU care, we can implement directed changes targeting patient-identified concerns. Methodology A survey of 37 questions, based on a validated national survey of family satisfaction with ICU care (FSICU-24) was put together addressing issues ranging from emotional care to technical aspects of the SICU experience on a whole. It is administered to the family member who was most involved in the patients care in the Surgical ICU after transfer out of the ICU. All the data is securely maintained and analyzed through a REDCap database for the purposes of quality improvement. Observations Over the initial weeks of administration, a few observations for improvement opportunities have been repetitive. Many patients and family members highly appreciate their attending doctors but cannot keep track of their names with the many teams of doctors. Having a time frame in which the doctor would arrive on rounds such that the family member can be present would be very helpful. The family members of patients sometimes feel uncared for in the SICU. Many would appreciate having someone show them the cafeteria or simply ask them if they need anything in particular. The plan of the day sheet (checklist as well as a list of the medical plan the team intends to follow) is often not given and/or explained to the patient and their family. Conclusion The following changes will be considered for feasibility of implementation. Surveys will be continuously administered in order to observe the effect the implemented changes have had. For example, changes might include the nurses explaining the plan of the day sheet to the patient and their family after the doctor has stated the plan of care; having picture cards of doctors with their name and photo would help patients and families better identify their caregivers; a volunteer could go around the ICU once a day and ask if the family has any needs. The expectation is to see improved patient and family satisfaction in those selected areas.

Nicole Burnett, Peter Mo, Naresh Rajan, Randy Madsen, Ram Gouripeddi, and Julio Facelli, University of Utah Life Sciences Background In the Salt Lake Valley there are three permanent Environmental Protection Agency (EPA) certified air quality monitoring stations that intake air samples and produce results of the air quality in the proximity of the monitor station. Due to the fact that the monitors only represent a small area of the 500- square-mile Salt Lake Valley, there are spatial gaps when using the air quality monitoring data for epidemiological studies. Since for certain studies health researchers may require a higher resolution spatiotemporal air quality grid [1], we need to devise new approaches to provide air quality data that could meet the epidemiological studies requirements. Research Methodology Modeled air quality data available from the EPA, has higher spatial and temporal resolution than data from monitoring stations, but it needs experimental validation and uncertainty quantification (UQ) in the Salt Lake Valley. We can achieve these validation and UQ goals through statistical comparisons of measured air quality data at the same location and time as the modeled air quality data. The air quality model that we primarily used is the one that the EPA has developed for the Centers for Disease Control and Prevention’s (CDC) National Environmental Public Health Tracking Network [2]. This is a model that uses a Hierarchical Bayesian Space Time Modeling approach [3]. This model was validated on the east coast of the United States so it is unknown how effective is in taking into account the terrain of the Salt Lake Valley. Modeled PM2.5 data in a 12×12 kilometer continuous grid resolution for the years 2007 and 2008 were compared against measured data of the same timeframe and location. The measured data was obtained from EPA’s Air Quality System (AQS) Datamart [4]. The statistical comparisons performed using these the two data sets were done using daily and monthly PM2.5 averages for the years 2007 and 2008 using MySQL, MATLAB and R. Conclusion & Significance We have developed a prototype for comparing and validating modeled air quality data against measured air quality data for the Salt Lake Valley. We found the modeled data fits the measured data fairly well. We will expand our work by developing a validating framework that will include a library of data modeling algorithms such as, The Complex Terrain Dispersion Model Plus Algorithms for Unstable Situations (CTDMPLUS) [5] and Yanosky’s [6], which could be selected by the user. The framework will be developed using OpenFurther, and then integrated with biomedical data [7]. The framework will be integrated into the PRISMS project [8] as part of the informatics infrastructure for studying the effects of air quality on pediatric asthma. References M. Z. Al-Hamdan, W. L. Crosson, A. S. Limaye, D. L. Rickman, D. A. Quattrochi, M. G. Estes, J. R. Qualters, A. H. Sinclair, D. D. Tolsma, K. A. Adeniyi, and A. S. Niskar, “Methods for Characterizing Fine Particulate Matter Using Ground Observations and Remotely Sensed Data: Potential Use for Environmental Public Health Surveillance,” J. Air Waste Manag. Assoc., vol. 59, no. 7, pp. 865-881, Jul. 2009. “Air Quality Data for the CDC National EPHT Network | Human Exposure and Atmospheric Sciences | US EPA.” [Online]. Available: http://www.epa.gov/heasd/research/cdc.html. [Accessed: 18-Sep-2014]. N. J. McMillan, D. M. Holland, M. Morara, and J. Feng, “Combining numerical model output and particulate data using Bayesian space-time modeling,” Environmetrics, p. n/a-n/a, 2009. “AirData | US Environmental Protection Agency.” [Online]. Available: http://www3.epa.gov/airdata/index.html. [Accessed: 20-Oct-2015]. Environmental Protection Agency, “Revision to the Guideline on Air Quality Models: Adoption of a Preferred General Purpose.” Environmental Protection Agency, 09-Nov-2005. J. D. Yanosky, C. J. Paciorek, F. Laden, J. E. Hart, R. C. Puett, D. Liao, and H. H. Suh, “Spatio-temporal modeling of particulate air pollution in the conterminous United States using geographic and meteorological predictors,” Environ. Health, vol. 13, no. 1, p. 63, Aug. 2014. R. Gouripeddi, N. Sundar Rajan, R. Madsen, P. Warner, and J. C. Facelli, “Federating Air Quality Data with Clinical Data,” presented at the 2014 AMIA Annual Symposium Proceedings, 2014. “Pediatric Research Using Integrated Sensor Monitoring Systems | National Institute of Biomedical Imaging and Bioengineering.” [Online]. Available: http://www.nibib.nih.gov/research-funding/prisms. [Accessed: 20-Oct- 2015]. Acknowledgements Grants: UU Air Quality Program, U54EB021973, NCRR/NCATS UL1RR025764, 3UL1RR025764-02S2, AHRQ R01 HS019862, DHHS 1D1BRH20425, UU Research Foundation. CHPC at UU.

Stephanie Bartlett, Utah Valley University Life Sciences Our project investigates the phylogenetic relationships of the superfamily Ephemeroidea + Behningiidae. Found in waters worldwide, burrowing families are unique in that they have mandibular tusks that allow them to tunnel in the silt or gravel of riverbeds. Surprisingly, even without the mandibular tusks necessary for tunnel construction, the Behningiidae family is still found within these burrows as nymphs. Because the Behningiidae don’t have tusks, morphological research has lead to a phylogenetic classification of this family that our genomic investigation ultimately disputes. To begin this investigation, mayfly specimens were collected worldwide and prepared for DNA extraction. For each specimen genes were amplified via polymerase chain reaction and visualized on an agarose gel, before being sequenced and analyzed. The specific genes targeted for this analysis include; 12s mitochondrial rDNA, 16s mitochondrial rDNA, 18s nuclear rDNA, 28s nuclear rDNA, H3 nuclear protein coding, and CO1 mitochondrial protein coding. Datasets were supplemented with sequences acquired from Genbank. The ingroup consisted of approximately 30 samples. Phylogenetic relationships were estimated using Maximum Parsimony, Maximum Likelihood, and Baysian methods. We constructed phylogenetic relationships of burrowing mayflies using molecular DNA data analysis, when compared to morphological analysis we identified some important classification differences. As a result of our findings, we propose an alternative explanation for the evolution of mandibular tusks in burrowing mayflies (Ephemeroptera). The data support that burrowing mayflies first evolved tusks. Within the family Behningiidae, tusks were lost, while the burrowing lifestyle (i.e., living in a burrow to filter feed) was retained. This study represents the largest analysis to date for these insects and strongly supports the evolutionary trend of a gain and a loss of mandibular tusks during their evolution.

Jake Loveless, Utah Valley University Life Sciences Macroinvertebrate assemblage composition was assessed in two perennial streams, Pleasant Creek, and Sulphur Creek in Capitol Reef National Park, Utah, during the summer of 2014. Individuals collected were identified to genus. This information was used to compare assemblage composition, nutrient availability, and water quality between the two streams. Five collection trips were made (May 27, June 10,17, July 1,15). Sampling was conducted randomly using mesh kick nets with four samples being collected per site, per trip. Heavy rains caused flash flooding on July 15 prohibiting sampling, so a total of sixteen samples per site were taken. Samples were stored in 90% isopropyl alcohol until they could besorted and identified using a dissecting microscope. Water quality estimates were determined by taking the weighted average of the tolerance values, and the final classification was made using the family-level biotic index. Pleasant Creek showed the highest number of taxa present, fourteen, compared to six in Sulphur Creek. Both streams followed the predictions of the River Continuum Concept consisting of collector-dominated functional feeding groups, with Pleasant Creek showing a higher percentage of predators 35.7% to 16.6%. Tolerance values of collected taxa were used to estimate organic pollution indicating good water quality in Pleasant Creek, and fairly poor water quality in Sulphur Creek. The differences in the macroinvertebrate assemblage composition in this study were likely multi-causal. Stream size is a major factor influencing the structure of macroinvertebrate assemblages; in general, as stream size increases, more taxa are added. Pleasant Creek is a much larger stream with a well-established channel, while Sulphur Creek is shallow with depths never exceeding fifteen centimeters. The shallow channel of Sulphur Creek also made it prone to frequent flooding. Isolated floods have little impact on macroinvertebrate communities due to availability of aerial adults nearby to recolonize rapidly, however, frequent flooding may have long-term effects through extirpation of taxa with high mortality. The difference in water quality was the most surprising finding in this study. The fairly poor water quality of Sulphur Creek may also be a factor in the lack of genera found there. Organic pollution effects primary productivity in streams, and while primary productivity was not measured in this study, visual observations found very little algae or aquatic vegetation at Sulphur Creek.

Alysa Edwards, University of Utah Life Sciences Percutaneous devices (PDs) constitute foreign materials that penetrate through the protective skin barrier to provide connection between internal and external environments. It has been previously shown that the periprosthetic tissue at the PD-skin interface is under a continuous state of wound healing, which often results in epidermal downgrowth. This continuous downgrowth is detrimental to the long-term survival of these devices. To date, there are no effective methodologies available to either prevent or quantify the degree of epidermal downgrowth indicating a need to find effective markers to document the healing response around these devices. In this study, periprosthetic tissues from a previous pig- back study were subjected to two different evaluations: (1) standard histology (HandE) and (2) immunohistochemical staining (IHC). Healing responses around PDs made with different material types were examined using cytokeratin 6 and collagen 4 to determine the degree of wound healing and granulation tissue maturity. Varying exposure time and concentration of stains, staining procedures were optimized. The interfacial tissues were then analyzed using either a photo or a confocal microscope. Preliminary data (Figure 1) indicated that there were noticeable differences in the periprosthetic regions between the material types used. The IHC data confirmed that the periprosthetic tissue is a hyper cellular region with a high density of blood vessels (collagen 4) and migrating keratinocytes. This data further confirmed the morphological differences observed between implant types using standard histology. Continued analysis will quantify the amount of collagen 4 within the periprosthetic tissue using imageJ software. Semi-quantitative data from each implant type will then be compared to predict biocompatibility. This research has demonstrated that IHC staining could be a potential tool for understanding the healing cascades around the percutaneous device.

Derek Munday, University of Utah Life Sciences Cardiovascular diseases (CVD) are more prevalent in individuals with diet-induced obesity (DIO) and type two diabetes (T2DM). Examples of CVD include blood vessel dysfunction and systemic hypertension. Both of these pathologies are associated with a reduced ability of the inner lining of the blood vessel (the endothelium) to release a substance (nitric oxide) that causes the blood vessel to dilate. At present the mechanism whereby T2DM and DIO decrease the function of the enzyme (nitric oxide synthase; NOS) responsible for nitric oxide synthesis and release is unknown. Determining this mechanism is the current focus of our laboratory. Earlier we reported that the sphingolipid ceramide is elevated in cell models of lipotoxicity and in mice with DIO. Most recently in endothelial cells we showed that ceramide causes protein phosphatase 2A (PP2A) to bind directly with NOS which disrupts the interactions among proteins that are necessary for optimal NOS function. My overall project was concerned with determining whether this mechanism is operational in mice with DIO. Specifically, we tested the hypothesis that PP2A inhibition would preserve vascular protein- protein interactions required for optimal NOS enzyme function to an extent that arterial dysfunction and hypertension would not occur. Mice consumed a control (CON) or high fat (HF) diet for 12 weeks. During the last 2 weeks, cohorts of mice from each group were injected (IP) with saline (vehicle control) or the PP2A inhibitor LB1 (1.0 mg/kg/day). We observed that interactions among proteins required for optimal NOS enzyme function were disrupted in arteries from mice with DIO treated with saline but not with LB1. Furthermore, arterial dysfunction and hypertension existed in mice with DIO that received saline but not LB1. These results strongly suggest that PP2A activation contributes importantly to arterial dysfunction that exists in a pre-clinical model of DIO.

Jacob Welch, Brigham Young University Life Sciences Learning and memory occur due to adaptive brain changes in response to our environment. These changes are mediated by synaptic plasticity, particularly within the hippocampus. Plasticity can either strengthen or weaken synapses, known as long-term potentiation (LTP) or long-term depression (LTD) respectively. While many forms of plasticity are NMDA-dependent, recently endocannabinoids were identified to mediate several new forms of hippocampal synaptic plasticity through the CB1 and TRPV1 receptors. However, research has demonstrated a non-CB1/TRPV1-dependent endocannabinoid synaptic plasticity in the hippocampus. Several potential candidate receptors that bind the endocannabinoid anandamide have been identified. These are among the orphan G-protein coupled receptors (GPRs) whose distribution in the brain and/or function is less well known. GPR55 is of particular interest as it activates second messenger systems. Using quantitative RT-PCR, electrophysiological and memory behavioral tasks we examined hippocampal GPR55 expression and function. GPR55 is expressed in hippocampus of both rats and mice. Cellular expression is currently being examined and appears to be rare in interneurons and more likely expressed by pyramidal cells. Interestingly, application of the GPR55 agonist LPI (2 μM) to wild-type mice demonstrates a significant enhancement of LTP in brain slices. This LPI effect was not noted in GPR55 knock-out (KO) mice, which exhibit significantly (p < 0.05) smaller LTP (146%) than wildtype (WT) (181%). GPR55 also appears to increase release probability (Sylantyev et al., PNAS, 2013), denoting a presynaptic role. Paired-pulse ratios are now being analyzed between GPR55 KO and WT mice to confirm this finding; however we did not note a change in EPSCs in CA1 in response to 2μM LPI. These data suggest GPR55 is expressed and physiologically relevant in the hippocampus. Because enhanced LTP is usually associated with better memory performance in rodents, this provides a potential target to enhance the cellular mechanism associated with memory formation.

Anne Thomas and Richard Gill, Brigham Young University Life Sciences Providing water for expanding urban communities in the western United States is a growing concern for city planners and governmental agencies. Landscaping can claim up to 50% of the urban water budget but also has the most potential for water conservation. Landscape water use is highly variable, however, because of species-specific differences in tree water use and because of decision-making by city planners and residents in maintaining trees. The objective of this study is to improve our ability to predict urban forest water use by identifying differences in tree basal area and diversity between neighborhoods that have arisen at different periods of development in Heber Valley, Utah. We classified neighborhoods as established, exurban (rural housing), commercial, or new tract based on age, location, and lot size. We performed a stratified random survey with twenty lots in each category and collected diameter and species data for each tree in the lot. Some of the patterns we observed were easily anticipated, such as higher basal area per hectare in the older, established neighborhoods relative to newer tract housing. Surprisingly, the number of individual trees per hectare in tract and established neighborhoods is very similar. Perhaps of more interest is the low species richness of tract housing compared to exurban communities. Because exurban communities are being replaced by tract housing there is evidence that tree diversity will be lost. Another important aspect of community structure in urban forests is the ratio of conifers to broadleaf trees because of fundamental differences in water use patterns. Conifers comprised twenty-five percent of the basal area in exurban and thirty-five percent in established neighborhoods, as opposed to five percent in tract. Our data suggest that tree diversity is likely to decrease while water demand is likely to increase with changes in urban forests in the coming decade.

Katherine Harris, Brady Evans and Thomas Andros, Brigham Young University Life Sciences Metabolic diseases, such as obesity and diabetes, have become endemic and the need for better treatments is rising. Mutations in PAS kinase, a recently discovered sensory kinase, have been shown to cause Maturity Onset Diabetes of the Young (MODY) in humans (Semplici et al., 2011). In addition, PAS kinase deficient mice display many phenotypes related to diabetes including resistance to weight gain, insulin insensitivity and triglyceride accumulation in response to a high-fat diet (Hao et al., 2007). Despite its importance in metabolism, little is known about the regulation of PAS kinase. PAS kinase consists of a sensory PAS domain that binds to and inhibits a protein kinase domain (Amezcua et al, 2002). We are currently engaged in several yeast genetic screens which will allow identification of regions in the full length PAS kinase that are essential for activation or for binding its substrates. The first screen is based on the finding that PAS kinase overexpression rescues a temperature-sensitive mutation in Tor2, the tor2(ts). We have isolated both point mutations and truncations in PAS kinase which alleviate the tor2(ts), suggesting they are hyperactive alleles. These mutations identify novel regions involved in PAS kinase regulation. Our second screen uses the yeast 2-hybrid to select for both point mutations and truncations that increase the ability of PAS kinase to bind its substrates. These mutations will help identify key regions of PAS kinase utilized in substrate recognition. Finally, we have identified regions of PAS kinase that are well-conserved throughout evolution and will compare these regions with the regions affected by our mutations. This study will be the first reported mutagenic analysis of PAS kinase. Analysis of these specific genetic regions will help elucidate the molecular mechanisms involved in the regulation and function of PAS kinase, a key player in the development of metabolic disease.

Scott Newton, Brigham Young University Life Sciences The ventral tegmental area (VTA) is known to controls the processing of rewarding and addictive behaviors. The VTA contains dopamine (DA) cells, which release DA to downstream targets in response to rewarding stimuli, and GABA cells, which modulate DA cell activity. Therefore, both cell types are involved in associative reward learning. Synaptic plasticity plays an important role in adaptive reward signaling within the VTA. Endocannabinoids mediate or modulate synaptic plasticity at synapses within the reward circuit. However, the source of endocannabinoids within the VTA is not well understood. Therefore, our goal was to describe the distribution of endocannabinoid biosynthetic enzyme mRNA within VTA neurons. We extracted single VTA neurons via whole cell patch clamp and used single-cell real-time quantitative PCR to identify DA and GABA neurons based on mRNA expression of cell-type specific targets. Additionally, electrophysiological properties such as action potential frequency and sag potential amplitude were examined between the two cell types. Concurrent with established observations, slower firing frequencies were observed in DAergic neurons, however overlap was identified between these two cell types. VTA neurons were then probed for endocannabinoid/ biosynthetic enzyme mRNA, such as N-acyl-phosphatidylethanolamine-specific phospholipase D (NAPE- PLD), diacylglycerol lipase α (DAGLα), and 12-lipoxygenase. We also probed for type I metabotropic glutamate receptor (mGluR) mRNA, as endocannabinoid synthesis requires mGluR activation. Our data demonstrate that endocannabinoid biosynthetic enzyme mRNA is expressed in both DAergic and GABAergic cells with concurrent expression of type I mGluRs. Next, to ensure mRNA expression was representative of protein content, slices were stained using immunohistochemistry for GAD67, DAGLα, NAPE-PLD and type I mGluRs. Positive labeling for these targets was observed in VTA neurons, supporting our RT-PCR results. Collectively, these data suggest DAergic and GABAergic cells of the VTA have the capability to produce endocannabinoids and potentially alter synaptic plasticity involved in reward and addiction.

Alex Hyer, University of Utah Life Sciences Geochemical reactions associated with the process known as serpentinization can generate copious quantities of organic carbon and chemical energy that support life, but these reactions also greatly increase the pH of serpentinization sites. High-pH environments hinder ATP production because the low proton concentrations at high pH can result in a reversed proton gradient across cell membranes. Organisms present at serpentinization sites are not well-characterized, and adaptations to their alkaliphilic lifestyle are unknown. Bacillus pseudofirmus OF4 was not isolated from serpentinization sites, but it is a well- studied alkaliphile that has several distinct adaptations for ATP production in high pH environments. Its sequence for ATP synthase, the molecular motor responsible for ATP generation in the cell, contains several distinctive mutations including an AXAXAXA motif that increases the C-ring stoichiometry, a P51XXE54XXP57 motif that creates a distinct kink in the outer helix of the C-subunit, and a V21 mutation granting rotamer freedom to E54. In this study, we search for similar mutations in metagenomic databases containing DNA sequences collected from serpentinization sites. Analysis of alignments from the programs BLAST and Exonerate indicate that V21 is present in several alignments, but P51 is found in only a single alignment from serpentinization sites. The AXAXAXA motif is non- existent in our databases, but the more common GXGXGXG motif is present with alanine replacements occurring periodically. In general, these results indicate that our DNA sequences from high-pH sites of serpentinization are more reminiscent of typical neutrophiles than of the alkaliphile B. pseudofirmus OF4. Therefore, serpentinization- adapted organisms may harbor other adaptations to high pH such as local pmf regulation. Future work will test for quantitative correlations between environmental pH measurements and the incidence of mutations in ATP synthase in order to identify novel adaptations to high pH in serpentinization-driven ecosystems.

Matthew Durrant and Shuqing Xu, Brigham Young University Life Sciences Researchers at the Max Planck Institute for Chemical Ecology have recently sequenced the Nicotiana attenuata (wild tobacco) genome. This genome affords researchers new opportunities to understand the evolution of this organism. One method for analyzing the evolution of specific genes in a given genome is referred to as phylostratigraphy, which makes use of large-scale BLAST sequence similarity searches. I designed a pipeline using the python programming language that implements a phylostratigraphic analysis to estimate the evolutionary age of all ~35,000 genes belonging to N. attenuata. By analyzing the large amounts of data produced by this BLAST search, each gene was assigned an estimated age through comparing the taxonomies of all other organisms that share similar protein sequences. This effectively answers the question “How old is this gene?” for each gene in the entire N. attenuata genome. Previous studies have produced microarray data that tracks the transcriptomic response of N. attenuata to an herbivore attack. By comparing the newly gathered gene evolutionary age information with this previously gathered microarray data, several new insights into the molecular signaling pathways of N. attenuata were made. It was found that at 1 hour following an herbivore attack, the transcriptome of N. attenuata is evolutionarily young, suggesting that the initial response to an herbivore attack recruits genes that have evolved more recently in the organism’s evolutionary history. At 5 hours after attack, however, there is a distinct decrease in the overall age of the N. attenuata transcriptome, suggesting that the organism is recruiting more ancient genes that are used to reconfigure the transcriptome of the organism. Beyond 5 hours, the transcriptome is once again relatively young, and it is clear that it has indeed been reconfigured to provide a more herbivore-specific defense response. This demonstrates a novel, evolutionary approach to analyzing signaling pathways in plants.