DIFFERENTIAL GENE EXPRESSION AND MOLECULAR MECHANISMS ASSOCIATED WITH DEVELOPMENT OF PALE, SOFT AND EXUDATIVE (PSE) TURKEY MEAT By Yuwares Malila A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Food Science – Doctor of Philosophy 2013 ABSTRACT DIFFERENTIAL GENE EXPRESSION AND MOLECULAR MECHANISMS ASSOCIATED WITH DEVELOPMENT OF PALE, SOFT AND EXUDATIVE (PSE) TURKEY MEAT By Yuwares Malila The success of turkey breeding has coincided with an increased incidence of a meat quality defect known as pale, soft and exudative (PSE) meat. Application of molecular-based approaches such as genetic markers for animal selection or pathway intervention to prevent development of this meat defect have been suggested as a potential long-term solution. However, molecular mechanisms associated with this alteration remain unclear. The overall goal of this study was to obtain better understanding of molecular mechanisms underlying development of PSE turkey. The study comprised two specific aims: 1) to assess global differential gene expression between normal and PSE turkey; and 2) to confirm differences between normal and PSE meat samples at the protein level of a candidate gene selected from aim 1. Turkey breast muscle samples were collected from 22wk randombred control line (RBC2) and 16wk commercial (COMM) turkeys. Breast samples were classified as normal or PSE based on marinade uptake (high = normal, low = PSE). Total RNA was isolated from muscle samples with the highest (normal, n = 6) and the lowest (PSE, n = 6) marinade uptake. Transcriptome analyses were conducted using two platforms: the turkey skeletal muscle long oligonucleotide microarray, and deep transcriptome sequencing with an Illumina Genome Analyzer IIX (RNA-Seq). The microarray study of RBC2 samples revealed 49 differentially expressed transcripts (false discovery rate, FDR < 0.1). Genes selected for pathway analysis were determined using two criteria: fold change ranking (FC < -1.66, FC > 1.66) and FDR < 0.35. The calcium signaling pathway was highlighted as the top canonical pathway. In addition, changes in expression of genes in the actin cytoskeleton signaling pathway suggested altered structures of actin filaments that may affect strength and flexibility of muscle cells. In RNA-Seq analysis, four RNA samples for each of the extreme normal and PSE characteristics from the RBC2 line were sequenced (n = 4). Pathway analysis of 494 differentially expressed transcripts (FDR < 0.05) identified by RNA-Seq confirmed previously suggested changes in calcium homeostasis and organization of actin cytoskeleton. Pyruvate dehydrogenase kinase isozyme 4 (PDK4), which regulates glucose oxidation, showed substantial decreased expression with both microarray (FC = 25.9) and RNA-Seq (FC = -14.1); thus, this gene was chosen as a candidate gene for further evaluation. The protein abundance of PDK4 was significantly decreased (FC = -3.4, P < 0.001) in PSE samples (n = 6) of the RBC2 line. Reduced expression of PDK4 at both transcriptional (FC = -12.8, P < 0.05) and translational levels (FC = 2.8, P < 0.001) was also observed in PSE turkey of the COMM line (n = 6), supporting the biological relevance of PDK4 suppression in the development of PSE turkey, and also suggesting that the mechanism responsible for the decreased PDK4 in RBC2 turkey subpopulations has been maintained in a commercial line. By identifying several candidate genes including PDK4, this study lays the foundation for future studies aimed at defining the mechanisms of development of PSE turkey. ACKNOWLEDGEMENTS I would like to express sincere gratitude to my advisor, Dr Gale Strasburg, for his valuable guidance and tremendous support throughout my program in both my professional and personal development. I wish to extend my gratitude to the members of my guidance committee members, Dr. Alden Booren, Dr. John Linz, Dr. Catherine Ernst and Dr. Ike Kang, for their constructive comments and suggestions for the successful completion of my research. I also would like to show my appreciation to Dr. Rob Tempelman for performing microarray statistical analysis and providing professional advice during the study of this technique, and Kevin Carr for operating RNA-Seq bioinformatics and data analysis. Dr. Hui-Ren Zhou participated in total RNA isolation and crude protein preparation. My appreciation is also extended to Dr. Kent Reed (University of Minnesota) and Dr. Sandra Velleman (Ohio State University) for their critical reviews on my manuscripts, and Dr. Kelly Sporer for her endless patience during lab training and valuable suggestions throughout my Ph.D. program. I would like to give recognition to the staff of Department of Food Science and Human Nutrition, particularly Ms. Deb Klein, and of the 3rd floor of Anthony Hall, Ms. Nancy Ramney and Ms. Nicole Hall for their friendly assistance. The study in this dissertation is a sub-aim investigation in the project funded by National Research Initiative Grant no. 2005-35604-15628 to G.M.S., C.W.E., K.M.R., S.G.V. and R.J.T and Grant no. 2005-35503-16348 to G.M.S., ! iv! W.C., J.R.L. and A.M.B. from the USDA National Institute of Food and Agriculture. I am also thankful for financial support throughout my Ph.D. program from Royal Thai Government. For all of my friends, JoAnn and Larry Woods, Josephine Wee, Dr. Brenna Flannery, Dr. Xiao Pan, Katie Horton, Sunisa Roidoung and Kishore Kalapurackal Janardhanan in particular, thank you so much for your neverending support and friendship as well as making my 5 years in East Lansing full of memories, and it was not miserable memories. Finally, I am grateful to my family, who spared my absence for the time I spent in Michigan State University, for their infinite love and support throughout the program. ! v! TABLE OF CONTENTS LIST OF TABLES............................................................................................. viii LIST OF FIGURES .......................................................................................... ix CHAPTER 1 INTRODUCTION ............................................................................................ CHAPTER 2 DIFFERENTIAL GENE EXPRESSION BETWEEN NORMAL AND PALE, SOFT AND EXUDATIVE TURKEY MEAT IDENTIFIED BY MICROARRAY ANALYSIS...………………….….……………………………..... ABSTRACT ........................................................................................... 2.1 Background……………………………………………………………... 2.2 Materials and methods………………………………………………… Experimental animals and sample collection.................................. RNA isolation.................................................................................. Microarray experimental design..................................................... RNA amplification and microarray hybridizations........................... Microarray statistical analysis and gene annotation....................... Confirmation of expression patterns............................................... Pathway analysis............................................................................ 2.3 Results…………………………………………………………………... Differential gene expression between normal and PSE turkey skeletal muscle..................................................................... Confirmation of microarray results by qPCR................................... Functional and pathway analysis using Ingenuity Pathways Analysis (IPA)................................................................................. 2.4 Discussion………………………………………………………………. Calcium signaling pathway.............................................................. RhoA signaling and actin cytoskeleton signaling pathway.............. Postmortem oxidative metabolism and PDK4................................. 2.5 Conclusions……………………………………………………………... CHAPTER 3 DEEP TRANSCRIPTOME SEQUENCING REVEALS DIFFERENCES IN GLOBAL GENE EXPRESSION BETWEEN NORMAL AND PALE, SOFT AND EXUDATIVE TURKEY MEAT………....................................…………..... ABSTRACT ........................................................................................... 3.1 Background……………………………………………………………... 3.2 Materials and methods………………………………………………… Sample Information......................................................................... RNA transcriptome sequencing analysis........................................ Confirmation of expression patterns............................................... ! vi! 1 8 8 9 11 11 12 13 13 14 15 16 18 18 23 25 27 29 34 37 40 41 41 42 45 45 46 47 Functional and pathway analysis.................................................... Microsomal membrane preparation................................................ Protein electrophoresis................................................................... Immunoblot..................................................................................... 3.3 Results…………………………………………………………………... Differential gene expression between normal and PSE turkey skeletal muscle revealed by RNA-Seq............................................ Confirmation of gene expression differences by qPCR technique......................................................................................... Functional and pathway analysis.................................................... Immunoblot of PDK4 protein........................................................... 3.4 Discussion………………………………………………………………. 3.5 Conclusions……………………………………………………………... CHAPTER 4 EXPRESSION OF PYRUVATE DEHYDROGENASE KINASE ISOZYME 4 IN PALE, SOFT AND EXUDATIVE MEAT FROM RANDOMBRED AND MODERN COMMERCIAL TURKEYS.....................…………….............……... ABSTRACT ........................................................................................... 4.1 Background……………………………………………………………... 4.2 Materials and methods………………………………………………… Sample information......................................................................... RNA Isolation and quantitative real-time PCR (qPCR)................... Microsomal membrane preparation................................................ Protein electrophoresis and immunoblot assay.............................. Image analysis................................................................................ Statistical analysis........................................................................... 4.3 Results…………………………………………………………………... Meat quality indices of normal and PSE turkey meat..................... Expression of PDK4 at transcriptional and translational levels in turkey meat................................................................................. 4.4 Discussion………………………………………………………………. 4.5 Conclusions……………………………………………………………... 48 50 51 51 53 53 58 60 64 66 72 74 74 75 77 77 78 79 79 80 81 81 81 83 85 86 CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTION………………………………….... 91 APPENDICES……………………………………………..……………………….. 96 Appendix A: Classification of pale, soft and exudative turkey meat..... 97 Appendix B: Differentially expressed transcripts identified by RNA-Seq analysis (FDR<0.05)........................................ 104 REFERENCES……………………..……………………..……………………….. 122 ! vii! LIST OF TABLES Table 2.1 Primer information for genes chosen for confirmation of expression using qPCR……………………………………………………………….. 17 Table 2.2 Number of differentially expressed transcripts between PSE and normal turkey skeletal muscle at different FDR significance levels… 19 Table 2.3 Differentially expressed transcripts between PSE and normal turkey skeletal muscle (FDR < 0.1)…………………………………………….. 19 Table 2.4 Top ten down-regulated (A) and top ten up-regulated (B) genes (based on fold change alone) in PSE relative to normal turkey skeletal muscle revealed by the TSKMLO microarray……………….. 22 Table 2.5 Top canonical pathways associated with development of PSE turkey……………………………………………………………………… 26 Table 3.1 Meat quality indices of the meat samples utilized in the RNA-Seq study……………………………………………………….. 46 Table 3.2 Primer information for genes chosen for confirmation of expression identified in the RNA-Seq analysis using qPCR……………………….. 49 Table 3.3 Number of loci classified within different Cufflink class code…………. 54 Table 3.4 Number of differentially expressed transcripts at different significance levels………………………………………………………… 54 Table 3.5 Potential downstream activities of the top altered biological functions associated with differential gene expression in PSE turkey…………………………………………………..…………….. 63 Table 4.1 Meat quality indices of the meat samples utilized in the study of PDK4 expression in RBC2 and COMM turkey………………………… 82 Table A.1 Cutoff values used in previous literature for classifying poultry PSE meat………………………………………………………………….. 99 Table B.1 Differentially expressed transcript identified in RNA-Seq study (FDR<0.05)……………………………………………………………...… 104 ! viii! LIST OF FIGURES Figure 2.1 Confirmation of gene expression analyzed by microarray using qPCR………………………………………………………………………. 24 Figure 2.2 Schematic diagram of the calcium signaling pathway associated with development of PSE turkey.........…………………………………. 31 Figure 2.3 Schematic diagram of RhoA and Actin cytoskeleton signaling pathways associated with development of PSE turkey…………........ 35 Figure 3.1. Distribution (%) of differentially expressed transcripts (FDR<0.05) at different fold change (FC) ranges. ………………………………….. 56 Figure 3.2 Schematic diagram showing localization and associated functions of the most up-regulated and down-regulated genes identified by RNA-Seq. ……………………………………………………………... 57 Figure 3.3 Confirmation of gene expression analyzed by RNA-Seq using qPCR……………………………………………………………….……… 59 Figure 3.4 Diagram of the calcium signaling pathway associated with development of PSE turkey.…………………….........………………… 61 Figure 3.5. Diagram indicating overlapping canonical pathways that regulate actin cytoskeleton and cell motility, and their associated differentially expressed genes (FDR<0.05).…………………………… 62 Figure 3.6 Expression of PDK4 protein in turkey breast muscle.………………… 65 Figure 3.7 Schematic diagram representing mitochondrial oxidative metabolism altered by decreased expression of PDK4 in PSE turkey.........…………………………………………………………. 70 Figure 4.1 Relative mRNA abundance of PDK4 between normal and PSE turkey as determined using qPCR.…………………………………….. 84 Figure 4.2 Difference in abundance of PDK4 protein between normal and PSE turkey.……………………………………………………………….. 85 Figure A.1 Color distribution of turkey breast meat.………………………………..101 Figure A.2 Correlations between lightness (L*) and meat quality indices (percent marinade uptake or percent cook loss) of turkey breasts.…………………………………………………….……… 102 ! ix! CHAPTER 1 INTRODUCTION For the past several decades, the meat consumption pattern in the US has shifted from predominantly red meat to white meat, as many consumers prefer meat and meat products that are low in both fat and calories. This trend is evidenced by the decrease in average annual consumption of red meat (beef and pork) from 134 lb per person in 1965 to 104 lb in 2011, and the increase in consumption of poultry from 40 lb per person to 100 lb in the same period of time (US Department of Agriculture, 2013). Turkey breast meat has become a preferred source of white meat. Due to its low fat content and high quality protein as well as its relatively inexpensive price, turkey and its processed products have become common daily food choices of consumers. In the US alone, average per capita annual consumption of turkey has increased from 8.7 lbs in 1978 to 15.7 lbs in 1988, and has remained steadily high at approximate 16 lbs from 1989 to 2012 (US Department of Agriculture, 2013). In response to an increase in consumer demand, turkey production has steadily risen. According to the Food and Agriculture Organization of the United Nations (2013), 1.5 billion turkeys were raised, yielding 32 billion lbs meat during the 1970s, but in the 2000s, global production of turkeys rose to 4.2 billion live birds producing 107 billion lbs of turkey meat. To meet this growing demand, turkeys have been intensively selected for rapid growth rate, large breast muscle mass and reduced abdominal fat. However, the success of breeding has coincided with an ! 1! increase in prevalence of a meat defect, described as pale, soft, and exudative (PSE) meat in turkeys (Dransfield and Sosnicki, 1999; Owens et al., 2000). The term PSE was originally used to describe pork with discoloration and poor water holding capacity (reviewed by Cassens, 2000). It is well recognized that development of both PSE pork and poultry are associated with rapid early-stage postmortem metabolism (Pietrizak et al., 1997; Le Bihan-Duval et al., 2003), leading to excessive lactic acid accumulation in the muscle while the carcass is still warm. The combination of high early postmortem acidity and heat initiates extensive denaturation of proteins; as a consequence, muscle fibers shrink and expel water in the form of drip that contributes to soft and flaccid texture of the PSE meat. The loss of myoglobin, the water-soluble red pigment in muscle, as exudate, causes the pale appearance. Denatured myofibrillar proteins present poor protein functionality characteristics, particularly water holding capacity and protein solubility. Such conditions detrimentally affect the quality of products from PSE meat. Packaged cuts surrounded by exudate are unappealing to customers. Most turkey is processed into value-added products and it is with these products that the PSE problem presents the greatest challenge. Because PSE meat possesses poor protein solubility properties, myofibrillar proteins are inadequately extracted during the manufacturing process, resulting in products with reduced cohesiveness and inconsistent texture (reviewed by Solomon et al., 1998). In addition to high cook loss resulting from low water holding capacity, products made from PSE meat exhibit low processing yield (Barbut et al., 2005; Bianchi et al., 2005). This defective product does not harmfully ! 2! affect consumer health, but it contains reduced level of water-soluble vitamins and minerals, and receives decreased consumer preference (Droval et al., 2012). Overall, the occurrence of PSE meat is economically troublesome for meat producers and processers. For the turkey industry, the problem of PSE meat has been a great challenge. In the US, PSE incidence causes an average $2 million loss per year to the processors (Owens et al., 2009). The prevalence of PSE turkey is a broad spectrum, but even within a single flock, incidence varies from 5-40% (Owens et al., 2000; Petracci et al., 2009; Zhu et al., 2012). Moreover, even small amounts of PSE meat as raw ingredients in a mixture undergoing processing can significantly affect quality of the final products. Several attempts have been made to improve the quality of meat products made from PSE turkey by changes in turkey management, through modification of transportation systems, and by processing interventions (reviewed by Barbut et al., 2008). However, these changes in practices have met with only moderate success. The use of food additives that increase water holding capacity and protein binding are promising but this approach generates more cost to the processers, and often does not yield an optimal product. An alternative strategy is to define the underlying mechanisms that make birds become susceptible to PSE development and exploit genetic markers associated with abnormalities in these pathways to select against PSE susceptibility. However, the molecular pathways underlying development of PSE turkey are still not completely understood. ! 3! Previous molecular studies regarding development of PSE turkey were developed based on similarities in accelerated early-stage postmortem metabolism in porcine and poultry PSE meat. It has been demonstrated that, in swine, the rapid postmortem metabolism is associated with a leak of the ryanodine receptor (RYR), the sarcoplasmic reticulum (SR) calcium channel protein, due to a single point 2+ mutation of RYR1 that causes [Ca ] overload inside the muscle cells (Mickelson et al., 1986; Mickelson et al., 1988a; Fujii et al., 1991; Otsu et al., 1994). However, in turkey, no mutation of RYR has yet been found (Chiang et al., 2007). Additionally, although a commercial DNA test for detecting RYR mutation in pigs has been developed and implemented, this test accounted for only 25-35% of PSE pork in commercial slaughter houses (reviewed by Barbut et al., 2008). This evidence suggests that development of the PSE phenotype is a multifactorial genetic problem involving not only RYR, but possibly many genes in multiple pathways. The long-term goal of this study is to minimize the occurrence of PSE turkey meat by using molecular-based knowledge to develop pathway intervention approaches. To accomplish this goal, complete understanding of molecular mechanisms associated with development of this turkey meat defect is required. The central hypothesis for this study is that there are differences at the molecular level between PSE and normal turkey providing complex interactions between 2+ intracellular Ca homeostasis and other metabolic pathways that lead to meat quality defects. To test the central hypothesis, the following specific aims are pursued. ! 4! Specific Aim 1: To identify global differences in transcript abundance between turkey breast muscle samples of birds classified as normal or PSE meat Previous studies focusing only on RYR were not able to explain development of the turkey PSE phenotype. The working hypothesis of this aim is that development of PSE turkey is associated with changes in expression of multiple genes responsible for numerous molecular pathways. Under this aim, global gene expression in breast muscle between normal and PSE turkey will be determined using two transcriptome analyses, 1) the turkey skeletal muscle long oligonucleotide (TSKMLO) microarray, and 2) the deep transcriptome sequencing technique (RNA-Seq). Transcriptome profiling techniques have been developed for simultaneous investigation of the expression of thousands of genes. Genes that are differentially expressed between treatments can then be clustered into networks based on their biological functions, which aids in identification of candidate pathways associated with the PSE phenotype. Each platform offers different advantages and disadvantages. The microarray method allows direct comparison of global gene expression between two treatments and reports the results as relative expression. The microarray is a well-established transcriptome profiling technique that is relatively inexpensive compared to RNA-Seq; thus, the microarray method allows us to conduct experiments with a larger number of biological replicates, that is more representative of the turkey population. In addition, the TSKMLO microarray, comprising 6000 70-mer oligonucleotides representing turkey skeletal muscle genes ! 5! (Sporer et al., 2011), is able to reveal differential expression of genes specific to turkey skeletal muscle. However, in recent years, microarray technology has been gradually superseded by RNA-Seq. Unlike hybridization-based microarray, RNA-Seq has the capability of quantifying actual mRNA abundance of the sample and determining genes beyond selected probes on the microarray (Roh et al., 2010). This platform also provides powerful dynamic range (Hitzemann et al., 2013) with high reproducibility and small technical variation (Marioni et al., 2008). Utilization of both techniques provides complementary results and facilitates greater understanding for development of PSE turkey. Specific Aim 2: To analyze abundance of a candidate protein encoded by a differentially expressed gene with the most biological relevance to development of PSE turkey The mRNA abundance determined by transcriptome analysis suggests changes in gene expression in response to biological stimuli and generally correlates with the proportion of encoded protein; however, this is not always the case (Preiss et al., 2003). For this aim, the working hypothesis is that protein abundance of a candidate gene, identified in aim 1, changes in the same direction with transcript difference. To accomplish this aim, a gene showing significant difference in expression between normal and PSE turkey from aim 1 will be selected as a candidate for determining protein abundance using protein immunoblot assay. The differentially expressed gene with the most apparent biological relevance to the ! 6! development of PSE meat will receive the highest priority for protein quantification in this aim. It is anticipated that results of this study will increase fundamental understanding of molecular mechanisms and pathways altered in PSE-susceptible turkeys resulting in reduced meat quality. Such information is foundational to identification of genetic markers for breeding selection and development of new intervention strategies to prevent development of PSE meat in susceptible birds. ! 7! CHAPTER 2 DIFFERENTIAL GENE EXPRESSION BETWEEN NORMAL AND PALE, SOFT AND EXUDATIVE TURKEY MEAT IDENTIFIED BY MICROARRAY ANALYSIS ABSTRACT In response to high consumer demand, turkeys have been intensively selected for rapid growth rate, and breast muscle mass and conformation. The success in breeding selection has coincided with an increasing incidence of pale, soft and exudative (PSE) meat defect, especially in response to heat stress. We hypothesized that the underlying mechanism responsible for the development of PSE meat arises from differences in expression of several critical genes. The objective of this study was to determine differential gene expression between normal and PSE turkey meat using a 6K turkey skeletal muscle long oligonucleotide microarray. Turkey breast meat samples were collected from the randombred control line at 22 weeks of age, and classified as normal or PSE primarily based on marinade uptake (high = normal, low = PSE). Total RNA was isolated from meat samples with the highest (normal, n = 6) and the lowest (PSE, n = 6) marinade uptake. Microarray data confirmation was conducted using quantitative real-time PCR. Selection of differentially expressed genes for pathway analysis was performed using a combination of fold change ranking (FC < -1.66, FC > 1.66) and false discovery rate (FDR < 0.35) as criteria. The calcium signaling pathway was highlighted as the top canonical pathway associated with differential gene expression between normal and PSE turkey. Dramatic downregulation of fast-twitch myosin heavy chain coupled with up-regulation of slow-twitch myosin and troponin C suggested a switch of skeletal muscle isoforms, which may ! 8 alter muscle fiber arrangement and formation of actin-myosin complexes. Changes in expression of genes in the actin cytoskeleton signaling pathway also suggest altered structures of actin filaments that may affect cell motility as well as strength and flexibility of muscle cells. Substantial down-regulation of pyruvate dehydrogenase kinase, isozyme 4 was observed in PSE samples suggesting altered regulation of the aerobic metabolic pathway in the birds that developed PSE meat defect. 2.1 Background Consumer demand for inexpensive food with low fat and high protein content has led to a tremendous growth of the poultry industry over the past several decades. Turkeys have become an attractive protein source because of their larger portion of lean meat compared to chicken. To meet increasing demand for turkey meat, birds have been intensively selected for rapid growth rate and breast muscle mass accretion and conformation (Barbut et al., 2008). However, the success of breeding has coincided with an increase in the prevalence of a significant quality defect known as pale, soft and exudative (PSE) meat (Dransfield and Sosnicki, 1999; Owens et al., 2000). Pale, Soft, Exudative meat was originally identified in pork with flaccid texture and unusually light color. Processed meat products made from either pork or poultry PSE meat are often of inferior quality (Aberle et al., 2001) and lower customer acceptability (Fernandez et al., 2002). This has been attributed to an extensive denaturation of proteins in PSE meat resulting in a loss of protein functionalities including solubility, water-holding capacity, and binding properties. Overall, the PSE ! 9 meat defect substantially lowers processing yields and causes significant economic loss to the poultry industry (Owens et al., 2009). In pigs, it is generally accepted that development of PSE pork associates with a rapid rate of postmortem anaerobic glycolysis, resulting in high carcass temperature and rapid decrease in pH, leading to protein denaturation. The 2+ accelerated postmortem glycolysis is associated with abnormal Ca 2+ muscle cells. A Ca homeostasis in leak via sarcoplasmic reticulum (SR) calcium release channel proteins (ryanodine receptors, RYR) in skeletal muscle cells, results from a single point mutation in RYR1 that changes the amino acid sequence from arginine at position 615 to cysteine (Mickelson et al., 1988b; Fuji et al., 1991; Otsu et al., 1994). In contrast, a hypersensitivity of the SR calcium release channels between different turkey lines has been suggested (Wang et al., 1999), but no mutation has yet been observed (Chiang et al., 2004). We hypothesized that some turkeys may be PSE-susceptible because of differences in the abundance of key proteins involved in 2+ regulation of intracellular [Ca ]. Recently, relative mRNA abundance of four major genes involved in Ca 2+ homeostasis in skeletal muscle cells between normal and PSE turkey breast meat has been determined (Sporer et al., 2012). With the onset of heat stress, the PSE meat showed a significant delay in the up-regulation of RYR 2+ isoforms, !RYR and "RYR, and of calsequestrin, the high-capacity, low-affinity Ca binding protein located in the lumen of the SR. Transcript abundance of the 2+ sarco/endoplasmic reticulum Ca -ATPase 1 (SERCA1) remained unchanged. This ! 10 previous study of Sporer et al. (2012) suggests a complex manifestation of changes in gene expression associated with development of PSE in turkey. It is possible that 2+ differential expression of unidentified genes other than SR Ca regulators may be revealed in PSE turkey. In addition, comparison of gene expression between normal and PSE turkey meat from birds not subjected to heat stress was not examined in the study of Sporer et al. (2012). This information is important to advance our fundamental comprehension at the transcriptional level regarding the development of PSE turkey meat. A turkey skeletal muscle long oligonucleotide (TSKMLO) microarray was constructed with the initial purpose of screening the skeletal muscle transcriptome for candidate genes critical for growth and development. The platform has been subsequently validated and utilized for studies of domestic turkey muscle biology (Sporer et al., 2011a; Sporer et al., 2011b; Nierobisz et al., 2012). Utilization of the TSKMLO platform enables a simultaneous investigation of the expression of thousands of skeletal muscle genes. The objective of the current study was to investigate differential expression in the turkey skeletal muscle transcriptome between normal and PSE meat using the TSKMLO microarray. 2.2 Materials and methods Experimental animals and sample collection Breast meat samples utilized in this study were obtained from the study of Chiang et al. (2008). Briefly, turkeys from the randombred control line 2 (RBC2), a ! 11 line representative of the commercial turkey of the late 1960s and maintained without selection pressure at The Ohio Agricultural Research and Development Center of The Ohio State University (Wooster, OH), were used in this study (Nestor et al., 1967; Nestor, 1977a; Nestor, 1977b). The birds were raised at the Michigan State University (MSU) Poultry farm (Chiang et al., 2008). All methods were approved by the Institutional Animal Care and Use Committee (IACUC; AUF#: 06/05-081-00). Turkeys at 22 weeks of age were harvested using standard industry practices in the MSU Meat Laboratory. Breast muscle samples from one side of each bird were collected immediately post-bleed, snap-frozen in liquid nitrogen, and stored at -80°C for later isolation of total RNA. Breast muscle samples from the opposite side were processed under commercial conditions. Breast muscle was classified after 24h as “normal,” primarily based on high marinade uptake, and secondarily on low cook loss. Conversely, “PSE” samples were grouped by low marinade uptake and high cook loss (Sporer et al., 2012). Six samples for each extreme of normal (average marinade uptake = 45.3%) and PSE (average marinade uptake = 17.2%) characteristics (n = 6) were used for microarray experiments. RNA isolation Total RNA was isolated from breast meat samples using Ambion TRI Reagent Solution (Applied Biosystems, Inc., Foster City, CA), and subsequently purified with the Qiagen RNeasy Mini spin column (Qiagen, Inc., Valencia, CA) according to manufacturer’s instructions. Quantity of total RNA was measured using a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Integrity of total RNA was confirmed using an Agilent 2100 Bioanalyzer (Santa Clara, CA). ! 12 Samples with an RNA Integrity Number (RIN) equal to or exceeding 8.0 (RIN = 10 is the best) were used for microarray and quantitative real-time PCR (qPCR). Microarray experimental design The 6K TSKMLO microarrays were used for transcriptome analysis of normal and PSE meat samples. Details of the array design are available at the National Center for Biotechnology Information’s Gene Expression Omnibus (NCBI GEO) with the platform accession GPL9788. Gene expression between normal and PSE turkey skeletal muscle samples was directly compared. Six biological replicates were run for each meat quality level (n = 6). Dye swapping was performed to minimize dye bias; i.e., in 3 arrays, the normal samples were labeled with Cy3 fluorescent dye (GE Healthcare, Piscataway, NJ) while PSE samples were labeled with Cy5; for the other 3 arrays, the dye assignments were reversed. A total of six arrays were utilized in this study. RNA amplification and microarray hybridizations Amplification of RNA samples, microarray preparation and microarray hybridization were performed as described elsewhere (Sporer et al., 2011a). Briefly, 2µg of total RNA was incubated with T7 Oligo(dT)18 as primer and reversetranscribed into cDNA followed by in vitro transcription for amplified RNA (aRNA) TM using an Amino Allyl MessageAmp II aRNA kit (Ambion, Austin, TX) according to the manufacturer’s instructions. For dye labeling, 10 µg of aRNA samples were dyecoupled with Cy3 or Cy5 fluorescent dye at room temperature in the dark for 1 h. ! 13 Dye-labeled aRNA was purified; then, 5 µg of each sample was fragmented into 60–200 nucleotide fragments using RNA Fragmentation Reagents (Ambion, Inc.) at 70°C for 15 min. The fragmented Cy3-labeled aRNA was equally mixed with its Cy5labeled counterpart. Hybridization of microarray slides was performed in a GeneTac TM Hybridization Station (Genomic Solutions, Ann Arbor, MI) for 18 h. Arrays were washed with a series of wash solutions and dried by centrifugation. Afterward, the arrays were scanned using a Molecular Devices GenePix 4000B scanner (Molecular Devices, Sunnyvale, CA), and image analysis was performed using Molecular Devices GenePix Pro 6.0 software. Array spot intensities were exported as GenePix Results (GPR) files for statistical analysis. Microarray statistical analysis and gene annotation Dye intensity bias was normalized using "normexp" background correction method based on Ritchie et al. (2007). Normalized data were described as log2 fluorescent intensities ratio (Cy5/Cy3) or M-value, and statistically analyzed with a linear model using LIMMA (Smyth, 2005). Specific hybridization was confirmed by monitoring fluorescence intensities of negative control and mismatched oligonucleotides (Sporer et al., 2011a). The microarray data were submitted to the NCBI GEO with GEO accession number GSE36660. The oligos were annotated using NCBI BLASTn (http://www.ncbi.nlm.nih.gov). ! 14 Confirmation of expression patterns Thirteen genes were selected for further analysis by qPCR to confirm the microarray gene expression results. Of 13 genes, 12 genes were chosen based on their large fold changes (FC; magnitude of FC indicates ratio of expression of particular gene in PSE relative to normal sample; negative FC indicates decreased expression of the gene in PSE samples or down-regulation; positive FC indicates increased expression in PSE samples or up-regulation) obtained from the microarray study. In addition, profilin (PFN) was selected as it showed statistical significance in expression between normal and PSE sample (estimated false discovery rates, FDR < 0.1). Primers (Table 2.1) were designed using Primer Express 3.0 software (Applied Biosystems) and synthesized by Operon Inc. (Huntsville, AL). The confirmation protocol was as described in Sporer et al. (2011a). Briefly, 5 µg of total RNA from the same samples utilized for microarray was reverse transcribed to cDNA using Superscript III (Invitrogen). The cDNA was purified using ethanol precipitation and quantified with a Nanodrop ND-1000 spectrophotometer. Reactions included 10 ng cDNA, 300 nM primer mix, and POWER SYBR Green Master Mix (Applied Biosystems) and were run in an ABI Prism 7900 Sequence Detection System (Applied Biosystems). Threshold cycle was analyzed using Sequence Detection Systems Software (SDS) version 2.3 (Applied Biosystems). Beta-actin (ACTB) was used as an endogenous control gene (Sporer et al., 2012). Relative expression of genes of interest in PSE samples relative to normal samples was ! 15 calculated using 2 -!!Ct method (Livak and Schmittgen, 2001). Student’s t-test was performed to evaluate significant difference (P < 0.05) in gene expression between PSE and normal samples. Pathway analysis To identify biological functions of the annotated genes, pathway analysis was performed using Ingenuity Pathway Analysis (IPA) (Ingenuity® Systems, www.ingenuity.com). Two criteria, FC and FDR, were used to select the genes for pathway analysis. Genes with FC < -1.66 or FC > 1.66 and FDR < 0.35 were uploaded to the IPA. Canonical pathways associated with focus genes were generated from the software. ! 16 Table 2.1 Primer information for genes chosen for confirmation of expression using qPCR Gene Symbol Primer ACTB ("-actin) ANKRD1 F: GTCCACCTTCCAGCAGATGTG R: CAATGGAGGGTCCGGATTC F: CGCCGATGCATGATGCT R: AATCAAAAGCCGGACCATTTT ATP1B4 F: TACCCTGGAAACGGCACATT R: TGTAGTTGACGTGCGTGAGCTT CA3 F: CAACCTGATGGTGTGGCTGTT R: TCTGGTTTGGGAGTTTTTCCA CCDC135 F: ATTACCTGGCACCTTTTCTTATTCA R: CGGAGGGCCTGCCTTTT COL6A1 F: TTCCATTGGTGCTCTTGCTATG R: TTTGGGATGATGGCGATACC CST3 F: GTGATCTCCAGAGCTGCGAAT R: ACAAAGGTGCATGTGGTATACTTAGC IGFN1 F: TCCCTGGTGATCTTTAGTGTTTCC R: CCTGAATCATTGGTGGCTTCA MYH4 F: AGCAGGCATTCACCCAACA R: GTCATGACGAGCAGACTGCAA MYOM1 F: TGAGCCAACTCCACAAGACAAA R: ATCAAATGCTTGCCCAGAAAGA PDK4 F: ATGAATGTCTGTAATAGTGCTTGCAA R: CATGTCTTCATTGTATGTTCTGCATATAC PFN2 F: CGGTCTTTCTGCCAGATCACT R: CGTCCAAACAGCGGCTTT RGS2 F: AGGCTCCCAAGGAGATAAACATT R: TCCTGGAGGTTCTGTGCTATCA TNNT3 F: CCCGTGCCTCAGTGATAACTAAA R: AGAAGAAAAGCAGCAGCAATAGC ! 17 Amp Length (bp) Amp Tm (°C) 71 79 56 81 70 80 62 79 67 79 79 78 67 80 71 79 104 82 100 76 90 74 65 82 64 79 68 78 2.3 Results Differential gene expression between normal and PSE turkey skeletal muscle Differential expression of 49 transcripts between PSE and normal turkey skeletal breast muscle was identified (FDR < 0.1). Seventeen transcripts were downregulated and 32 transcripts were up-regulated (Table 2.2). Among those transcripts, there were two corresponding to oligos annotated as alpha sarcoglycan (SGCA), three oligos annotated as myosin heavy chain isoform 2 (MYH2), two oligos annotated as nebulin (NEB), four oligos annotated as myosin heavy chain isoform 1 (MYH1), two oligos annotated as titin (TTN) and one unknown oligo (Table 2.3). The top 10 down- and up-regulated genes (based on FC only) are shown in Table 2.4. Down-regulated genes refer to the genes expressed lower in PSE relative to normal samples. Conversely, up-regulated genes are those expressed higher in PSE compared to normal sample. ! 18 Table 2.2 Number of differentially expressed transcripts between PSE and normal turkey skeletal muscle at different FDR significance levels Number of transcripts Downregulated Upregulated FDR < 0.05 FDR < 0.10 FDR < 0.15 FDR < 0.20 FDR < 0.25 FDR < 0.30 FDR < 0.35 10 17 32 47 74 101 131 19 32 43 57 71 97 113 Total 29 49 75 104 145 198 244 Table 2.3 Differentially expressed transcripts between PSE and normal turkey skeletal muscle (FDR < 0.1) GenBank Accession# Gene symbol DQ993255 MHC NM_001044651 ATP1B4 NM_204228.1 MYH2 Gene name Major histocompatibility complex, B locus ATPase, (Na+)/K+ transporting, beta 4 polypeptide FC FDR -14.6 0.003 -12.0 0.04 -7.1 0.07 -4.2 -3.6 -2.7 0.08 0.04 0.03 -2.4 -2.3 -2.3 0.08 0.04 0.08 -2.1 0.09 -2.0 0.01 XM_423397 Myosin, heavy chain 2, skeletal muscle RIN2 Ras and Rab interactor 2 CCDC135 Coiled-coil domain containing 135 LINGO1 Leucine rich repeat and Ig domain containing 1 STBD1 Starch binding domain 1 ATL2 Atlastin GTPase 2 ACSF3 Acyl-CoA synthetase family member 3 USP9X Ubiquitin specific peptidase 9, Xlinked PSMC6 Proteasome (prosome, macropain) 26S subunit, ATPase, 6 SELENBP1 Selenium binding protein 1 -2.0 0.04 NM_001079760 PFN2 -2.0 0.02 XM_419317.3 XM_003209725 NM_001024577 XM_429975 NM_001242311 XR_118375 XM_003203019 NM_001006494 ! 1 Profilin 2 19 Table 2.3 (cont’d) GenBank Accession# Gene symbol XM_003207146 HNRNPA2B1 XM_003209920 VPS4 XM_003213364 NM_205177 AQP3 CTSD XM_003643520 NM_001030899 SGCA THRAP3 XM_003643520 SGCA 2 2 AJ419877 XR_118187 PHKA1 XM_003203043 RPS6KA3 XM_003207155 TAX1BP1 NM_205096 HDLBP XM_415578 MYH2 1 BC067379 3 AB024330 NM_001013396 MYH1 XM_003204036 NM_001012945 ACTA1 DNAJA1 XM_001236426 ! NEB HDGF 4 FDR -1.9 0.08 -1.9 0.02 -1.8 0.02 1.8 1.9 Alpha sarcoglycan 1.9 0.09 0.01 0.08 Thyroid hormone receptor associated protein 3 2.0 0.04 Alpha sarcoglycan 2.4 0.02 18S rRNA gene Phosphorylase b kinase regulatory subunit alpha, skeletal muscle isoform Ribosomal protein S6 kinase, 90kDa, polypeptide 3 Tax1 (human T-cell leukemia virus type I) binding protein 1 2.4 2.5 0.08 0.07 2.5 0.049 2.6 0.08 High density lipoprotein binding protein (vigilin) Myosin, heavy chain 2, skeletal muscle, adult DR1-associated protein 1 (negative cofactor 2 alpha) 2.8 0.02 2.8 0.02 3.2 0.004 Nebulin EIF2S3 FC Vacuolar protein sortingassociated protein 4A Aquaporin-3 Cathepsin D NM_001006260 Gene name 3.3 0.02 Myosin heavy chain isoform 1, skeletal muscle, adult Actin, alpha skeletal muscle Heat shock protein 40 kDa homolog, subfamily A, member 1 Hepatoma-derived growth factor 3.7 0.004 3.7 3.8 0.08 0.02 3.8 0.02 Eukaryotic translation initiation factor 2, subunit 3 gamma Heterogeneous nuclear ribonucleoproteins A2/B1 20 Table 2.3 (cont’d) GenBank Accession# Gene symbol 1 XM_415578 MYH2 NM_204289 HSP90B1 XM_420181 PRPS1 NM_001031489 HBE1 XR_118264 TTN NM_001013396 MYH1 4 MYH1 AB024330 NEB MYOM1 3 5 TTN NM_001013396 MYH1 CA3 XM_418319 EF153719 FDR 3.9 0.08 3.9 0.08 3.9 0.08 4.1 4 4.2 0.02 0.08 Myosin heavy chain isoform 1, skeletal muscle, adult Clone AY006 28S ribosomal RNA gene, partial sequence. Myosin heavy chain isoform 1, skeletal muscle, adult 4.2 0.01 4.9 0.04 5.0 0.08 Nebulin 5.8 0.04 Myomesin 1 28S rRNA gene, clone GgLSU-1 7.4 7.4 Titin 4 NM_001013396 XR_118264 Myosin, heavy chain 2, skeletal muscle Heat shock protein 90kDa beta (Grp94), member 1 Phosphoribosyl pyrophosphate synthetase 1 Hemoglobin subunit beta FC Titin 5 DQ018757 NM_204959 FM165415 Gene name 8.0 0.02 0.04 0.08 Myosin heavy chain isoform 1 8.1 0.01 Carbonic Anhydrase 3 Mitochondrion, complete genome Unknown 14.6 18.9 0.06 0.04 -2.1 0.09 FC = Fold change Different oligos on the microarray but were annotated as myosin heavy chain isoform 2. 2 Different oligos on the microarray but were annotated as sarcoglycan. 1 3 4 Different oligos on the microarray but were annotated as nebulin. Different oligos on the microarray but were annotated as myosin heavy chain isoform 1. 5 Different oligos on the microarray but were annotated as titin. ! 21 Table 2.4 Top ten down-regulated (A) and top ten up-regulated (B) genes (based on fold change alone) in PSE relative to normal turkey skeletal muscle revealed by the TSKMLO microarray A. Down-regulated Genes Symbol MYH4 PDK4 MHC MB ATP1B4 ANKRD1 RSG2 IGFN1 TNNT3 CCDC135 Gene name Myosin, heavy chain 4, skeletal muscle Pyruvate dehydrogenase kinase, isozyme 4 Major histocompatibility complex, B locus Myoglobin ATPase, Na+/K+ transporting, beta 4 polypeptide Ankyrin repeat domain 1 (cardiac muscle) Regulator of G-protein signaling 2, 24kDa Immunoglobulin-like and fibronectin type 3 domain containing 1 Troponin T type 3 (skeletal, fast) Coiled-coil domain containing 135 FC -26.2 -25.9 -14.6 -13.8 -12.0 -6.8 -4.7 -4.2 FDR 0.11 0.33 0.003 0.25 0.04 0.15 0.41 0.14 -3.8 -3.6 0.41 0.04 FC FDR 0.06 0.01 B. Up-regulated Genes Symbol CA3 Gene name Carbonic anhydrase III, muscle specific MYH1 TTN MYOM1 Myosin, heavy chain 1, skeletal muscle, adult 14.6 8.1 Titin Myomesin 1 8.0 7.4 0.08 0.02 TNNC1 NEB COL6A1 CST3 Troponin C type 1 (skeletal muscle, slow) Nebulin Collagen, type VI, alpha 1 Cystatin C 7.0 5.8 5.7 4.3 0.34 0.04 0.25 0.37 1 1 Myosin, heavy chain 1, skeletal muscle, adult 4.2 < 0.01 MYH1 HBB Hemoglobin, beta 4.1 0.02 FC = Fold change FDR = False discovery rate 1 Different oligos on the microarray but were annotated as myosin heavy chain isoform 1 ! 22 Confirmation of microarray results by qPCR Confirmation of microarray analysis was performed by conducting qPCR analysis of 13 selected genes (Figure 2.1). The majority of the selected genes showed statistical significance in expression between normal and PSE turkey (P < 0.05) with a similar direction as observed in microarray analysis. Five genes including regulator of G-protein signaling 2 (RGS2), immunoglobulin-like and fibronectin type 3 domain containing 1 (IGFN1), cystatin C (CST3), troponin T type 3 (TNNT3), and myomesin 1 (MYOM1), did not show statistically significant differential expression in the qPCR experiments. It should be noted that, among those five genes, the FDR from the microarray study for RGS2, CST3 and TNNT3 exceeded 0.35. ! 23 Figure 2.1 Confirmation of gene expression analyzed by microarray using qPCR. Results are presented as relative expression or fold change for gene expression in PSE relative to normal samples. Bars below the origin indicate lower expression (down-regulation) of the gene in PSE samples; bars above the origin indicate higher expression (upregulation) in PSE samples. Statistical significance indicates change in expression between PSE and normal samples within each technique († FDR < 0.1 for microarray, * P < 0.05 for qPCR). ! 24 Functional and pathway analysis using Ingenuity Pathways Analysis (IPA) In order to perform pathway analysis, two parameters were utilized for gene selection from the microarray results. The primary criterion was FC < -1.66 or FC > 1.66. The second criterion was FDR < 0.35. Using these criteria, expression data from 174 transcripts were uploaded to IPA. According to the IPA knowledge base, 86 genes were recognized and mapped into canonical pathways (well-characterized cellular signaling pathways). Canonical pathways associated with development of PSE turkey are shown in Table 2.5. The calcium signaling pathway (Figure 2.2) was the first pathway suggested by the IPA and supports an association between 2+ development of PSE meat and abnormal Ca homeostasis. Ras homology family member A (RhoA) signaling and actin cytoskeleton signaling (Figure 2.3) were also identified. As suggested by IPA, changes in gene expression regarding development of PSE turkey also altered pathways related to development of muscle and muscle contraction. ! 25 Table 2.5 Top canonical pathways associated with development of PSE turkey Canonical Pathways 1. Calcium Signaling 2. RhoA Signaling 3. Hypoxia Signaling in the Cardiovascular System 4. Caveolar-mediated Endocytosis Signaling 5. Protein Ubiquitination Pathway 6. Nuclear factor-erythroid 2-related factor (NRF2)-mediated Oxidative Stress Response 7. Integrin Signaling 8. Actin Cytoskeleton Signaling FC = Fold change FDR = False discovery rate ! 26 Gene symbol MYH4 ATP2A1 TNNI2 ACTA1 TNNC1 MYLK2 PFN2 IGF1 ACTA1 P4HB UBE2D2 HSP90B1 FLNC ITGA7 ACTA1 USP9X PSMC6 UBE2D2 DNAJA1 HSP90B1 DNAJA4 AOX1 DNAJA1 ACTA1 MYLK2 ITGA7 CAPN3 ACTA1 MYH4 MYLK2 PFN2 ACTA1 FC FDR -26.1 -2.0 2.9 3.7 7.0 -2.1 -2.0 2.0 3.7 1.8 2.0 3.9 -2.0 -1.7 3.7 -2.1 -2.0 2.0 3.8 3.9 -2.9 -1.9 3.8 3.7 -2.1 -1.9 2.1 3.7 -26.2 -2.1 -2.0 3.7 0.11 0.25 0.16 0.08 0.34 0.28 0.02 0.28 0.08 0.28 0.17 0.08 0.25 0.25 0.08 0.09 0.02 0.17 0.02 0.08 0.23 0.25 0.02 0.08 0.28 0.25 0.18 0.08 0.11 0.28 0.02 0.08 2.4 Discussion Development of PSE in poultry poses one of the greatest challenges to the meat processing industry (Anthony, 1998; Petracci and Cavani, 2012; Samuel et al., 2012). The prevalence of PSE turkey meat can be as high as 50% in commercial plants depending on the flock, time of the year and other factors such as transportation (Barbut, 1998; Woelfel et al., 2002). However, even a small amount of PSE turkey entering the processing line can have significant negative effects on quality of the final meat products. Several investigators have attempted to define PSE meat by using objective measurements to establish a cut-off value that would separate PSE meat from normal (Barbut, 1998; Garcia et al., 2010; Eadmusik et al., 2011). One of the indicators frequently studied is carcass color, as this characteristic can be measured rapidly and is amendable to use on the processing line. However, this indicator is weakly correlated with percent marinade uptake (Chiang et al., 2008) and water holding capacity (Samuel et al., 2012). Thus, use of carcass color for on-line sorting of PSE from normal meat is problematic for the meat industry. The overall goal of this study is to gain a greater mechanistic understanding of the development of PSE in turkey so that new strategies can be developed to identify PSE-susceptible animals in the breeding population, and thus reduce the number of PSE birds entering the processing line. Turkeys from the RBC2 line, developed without selection pressure, were chosen for this study. Analysis of normal ! 27 and PSE breast samples from this line will serve as a basis for comparison with growth-selected turkey lines in future studies. 2+ In previous studies, delayed up-regulation of Ca regulatory proteins was observed in PSE turkey samples upon heat stress treatment (Sporer et al., 2012). The results suggested a greater complexity of the development of the PSE meat defect than a single gene mutation. It is possible that additional genes may be involved that have not yet been identified. Unlike the study of individual gene expression, the microarray technique enables simultaneous analysis of thousands of genes with the capability of revealing differential expression of unidentified genes and interaction among genes. In this study, the 6K TSKMLO microarray was utilized to identify relative transcript abundance between normal and PSE turkey meat samples. Considering the top 10 down- and up-regulated genes, several genes displayed large FC but were not significantly different based on statistical criteria (FDR < 0.1); examples include myosin heavy chain isoform 4 (MYH4, FC = -26.2, FDR = 0.11) and pyruvate dehydrogenase kinase, isozyme 4 (PDK4, FC = -25.9, FDR = 0.33). The MicroArray Quality Control (MAQC) project suggested a rationale for gene selection (Shi et al., 2008). After completing statistical analysis of numerous platforms, this group found that selection of genes based on a combination FC ranking and a less stringent P threshold improves reproducibility and specificity of microarray analyses. The more stringent the P cutoff utilized for gene selection, the less reproducible the list of differentially expressed genes (Shi et al., 2008). Based ! 28 on this project, a FC cutoff (FC < -1.66 or FC > 1.66) with a less stringent FDR (FDR < 0.35) was used in the present study to identify genes for more confidently performing pathway analysis. Using these expanded criteria, changes in gene expression for several of the selected genes, including MYH4 and PDK4, were confirmed by qPCR (P < 0.05). Pathway analysis revealed that several cellular signaling pathways are associated with the development of PSE turkey meat, with numerous genes associated with more than one pathway. For example, MYH4 was mapped into calcium signaling, actin cytoskeleton signaling, and protein kinase A signaling pathways. This result indicates interactions among molecular pathways associated with development of this meat defect. The potential roles of the calcium signaling, RhoA signaling and actin cytoskeleton signaling pathways in development of PSE meat are discussed below. Calcium signaling pathway The calcium signaling pathway (Figure 2.2) was the top pathway highlighted by the IPA. This pathway is of particular interest because several previous studies 2+ have implicated abnormal Ca homeostasis in the development of PSE meat. In pigs, it is widely accepted that development of PSE pork is associated with a single 2+ point mutation in RYR1, which results in abnormal Ca homeostasis in skeletal muscle of MH-susceptible pigs (Mickelson et al., 1988b; Fuji et al., 1991; Otsu et al., 2+ 1994). The rate of Ca ! release from SR in stress-susceptible pigs is about two 29 times greater than that of normal pigs (Mickelson and Louis, 1996). The high level of 2+ sarcoplasmic Ca postmortem activates muscle hypermetabolism and accelerates pH decline. With the combination of high acidity and high carcass temperature in the initial postmortem phase, proteins undergo denaturation, causing the PSE meat defect. Previously, Sporer et al. (2012) found a delay in up-regulation of !RYR and "RYR expression in PSE turkey compared with normal samples when the birds underwent heat stress. However, in the current study, there was no evidence of differential gene expression of RYR, either !-, or "- isoform, between normal and PSE turkey meat when the birds were not undergoing heat stress. This discrepancy may be due to the fact that the fold-change differences observed by qPCR were modest and thus not determined to be significant by the microarray method. These results agree with the study of Oda et al. (2009) who observed an unchanged expression of broiler !RYR, although they found decreased expression of "RYR. This may be due to the biological differences among species. ! 30 Figure 2.2 Schematic diagram of the calcium signaling pathway associated with development of PSE turkey. The pathway, suggested by the IPA, showed direct interaction among differentially expressed genes associated with regulation of Ca 2+ concentration between normal and PSE turkey skeletal muscle. (FC = fold change, negative FC = down-regulation in PSE samples, positive FC = up-regulation in PSE samples) ! 31 2+ Apart from RYR, the intracellular [Ca ] is also regulated by the SERCA, the SR-Ca 2+ pump. The SERCA isoform 1 is expressed exclusively in fast-twitch 2+ skeletal muscle and is encoded by the ATPase, Ca -transporting gene (ATP2A1) (Lytton et al., 1992). Mutation of ATP2A1 results in a reduction of SERCA1 activity and is associated with an exercise-induced impairment of skeletal muscle relaxation and severe cramps in humans (Odermatt et al., 2000) and cattle (Sacchetto et al., 2009). In this study, lower ATP2A1 expression was found in PSE samples (FC = -2.0). This may imply an overload of intracellular Ca 2+ causing a severe muscle contraction in the susceptible birds. It should be noted that expression of the gene encoding SERCA1 in PSE turkey remained unchanged with the onset of heat stress (Sporer et al., 2012); however, differential expression of this gene between normal and PSE turkey without heat stress treatment was not determined in the previous study. Pathway analysis clustered genes encoding myofibrillar proteins within the calcium signaling pathway. Expression of MYH4 encoding myosin heavy chain (MHC) isoform IIb (Tonge et al., 2010) dramatically decreased in PSE samples (FC = -26.1). In contrast, alpha actin (ACTA1) was up-regulated (FC = 3.7). Myosin and actin are major contractile proteins in skeletal muscle, which constitute approximately 45% and 20% of myofibrillar proteins, respectively (Aberle et al., 2001). By regulating myofibrillar assembly, changes in expression of genes encoding myosin and actin may affect myofibril accumulation and stability (Wells et al., 1996) and potentially lead to irregular organization of muscle fibers in PSE meat ! 32 as previously found in PSE pork (Laville et al., 2005; Obi et al., 2010) and PSE chicken (Wilhelm et al., 2010). Interestingly, while MYH4 (which encodes fast-twitch glycolytic MHCIIb) was down-regulated, MYH2 (which encodes fast-twitch oxidative glycolytic MHCIIa) and MYH1 (intermediate between type IIa and IIb MHCIIx) showed up-regulation. Because turkey breast muscle mainly comprises fast-twitch glycolytic muscle fibers (Rosser et al., 1996), it can be hypothesized that the change in transcript abundance of MYH gene family members in the PSE turkey may cause fast-to-slow muscle transformation which is associated with various functional changes at the muscle cell level as well as cytosol-regulating Ca 2+ dynamics (Kaprielian et al., 1991; Jakubiec-Puka et al., 1999; Pette and Vrbová, 1999; Pette and Staron, 2001; Tonge et al., 2010). However, previous reports on the relative proportion of myosin heavy chain isoforms in PSE meat have not been consistent (Ryu and Kim, 2006; Franck et al., 2007; Golding-Myers et al., 2010). Up-regulation of slow-muscle troponin C (TNNC1, FC = 7.0) corresponds to a shift of myosin isoforms, supporting the hypothesis of the fast-to-slow muscle type conversion. Biological properties of slow-muscle troponin C, including Ca 2+ binding, 2+ the Ca -bound conformation as well as interaction with troponin I, differ from that of the fast muscle isoform (Sia et al., 1997). In addition, fast-twitch troponin I (TNNI2, FC = 2.9) was up-regulated in PSE samples. The differential expression of the regulatory proteins in turkey skeletal muscle may result in change in the ratio of regulatory proteins after protein translation. Together, changes in expression of myofibrillar proteins can directly and indirectly alter interactions among the proteins ! 33 and their response to Ca 2+ flux, with the net effect being altered meat quality. RhoA signaling and actin cytoskeleton signaling pathway In this study, an alteration of pathways involved in regulation of arrangement of actin and actomyosin in PSE turkey meat was identified. These pathways include RhoA signaling and actin cytoskeleton signaling (Figure 2.3). RhoA, is a subtype of the Ras superfamily, a low-molecular-weight phosphoprotein family of GTPases (McClung et al., 2004) that links extracellular growth signals or intracellular stimuli to the assembly and organization of the actin cytoskeleton (Schmidt and Hall, 1998). Insulin-like growth hormone 1 (IGF1) was mapped by IPA into the RhoA signaling pathway. This protein, structurally similar to insulin, acts via either autocrine or paracrine mechanisms (McMurtry et al., 1997) and regulates tissue growth and development in various vertebrates (Jones and Clemmons, 1995), including turkeys (Bacon et al., 1993; Richards et al., 2005). Infusion of IGF-1 into chicken has been shown to affect protein synthesis (Conlon and Kita, 2002) and mediate protein degradation (Czerwinski et al., 1998; Tomas et al., 1998). Upregulation of IGF-1 in PSE turkey (FC = 2.0) may imply changes in protein turnover, including actin, thus impacting downstream actin cytoskeleton function in the defective meat. ! 34 Figure 2.3 Schematic diagram of RhoA and Actin cytoskeleton signaling pathways associated with development of PSE turkey. The pathway, suggested by IPA, shows interaction among differentially expressed genes associated with polymerization of actin filament and formation of actomyosin complex between normal and PSE turkey skeletal muscle. (FC = fold change, negative FC = down-regulation in PSE samples, positive FC = up-regulation in PSE samples) ! 35 The actin cytoskeleton is a polymer of actin monomers assembled together via condensation reaction (Schmidt and Hall, 1998). The primary function of the actin cytoskeleton in skeletal muscle is to tether structural components and maintain overall structural order of those components inside the cell, but in contrast to myofibrillar actin, the cytoskeleton is not directly involved in muscle contraction (Stromer, 1998). The actin cytoskeleton contributes strength and flexibility to accommodate changes in cell shape during muscle contraction. The cytoskeleton is also intimately involved in other cellular functions including cell division and transmembrane signaling (Schmidt and Hall, 1998). Organization of the actin cytoskeleton is determined by the turnover of actin filaments and actin-binding proteins which are modulated by either internal or by environmental signals (Sheterline and Sparrow, 1994). One of the important actinbinding proteins is profilin encoded by PFN2. Profilin catalyzes ATP/ADP exchange of actin monomers and inhibits the hydrolysis of ATP bound to monomeric actin thereby maintaining concentration of readily polymerizing ATP-bound actin (Sohn and Goldschmidt-Clermont, 1994; Dos Remedios et al., 2003). The protein also promotes formation of actin filaments by transporting the monomer to the growing end of the filament (Dos Remedios et al., 2003). In this study, transcript abundance of PFN2 was lower in PSE samples (FC = -2.0), which may affect the concentration of ATP-bound actin monomers and alter arrangement of actin filaments. Actin cytoskeleton can be formed along myosin filaments and generate force in the appropriate direction based on polarity of actin filaments. Myosin-based motility of actin cytoskeleton accounts for muscle contraction (when it interacts with muscle ! 36 myosin) or cellular morphogenic movement (when it interacts with non-muscle myosin) (Sheterline and Sparrow, 1994). Thus, altered assembly of actomyosin may affect cell motility. In this study, changes in expression of the gene encoding myosin light chain 2+ kinase (MYLK) may alter actomyosin assembly. The MYLK is a Ca /calmodulin dependent enzyme (Park et al., 2011) that phosphorylates the myosin regulatory 2+ light chain (RLC). When phosphorylated, the RLC mediates Ca sensitivity in myofilaments and promotes movement of myosin cross bridges away from the thick filament surface towards actin filaments (Stull et al., 2011) and affects skeletal muscle twitch potentiation (Manning and Stull, 1982). The MYLK enzyme encoded by the gene MYLK2 is expressed predominantly in fast-twitch skeletal muscle fibers (Zhi et al., 2005). In the study of Zhi et al. (2005), knockout of the MYLK2 gene decreased RLC phosphorylation in mouse skeletal muscle. Thus, change in expression of MYLK2 (FC = -2.1) in PSE meat suggests an alteration of stability of myosin molecules and the actomyosin complex. Postmortem oxidative metabolism and PDK4 The PDK4 gene was substantially down-regulated in PSE meat (FC = -25.9). This gene was not mapped by IPA into a specific canonical pathway. However, considering its function, PDK4 may be one of the key players regarding development of PSE turkey meat. This gene encodes the PDK4 enzyme, one of the regulators of oxidative metabolism in the mitochondria. The enzyme inhibits conversion of pyruvate into acetyl CoA by phosphorylating pyruvate dehydrogenase (Wynn et al., ! 37 2008). The dramatic down-regulation of the PDK4 gene suggests altered oxidative metabolism in PSE turkey meat. It can be hypothesized that, for normal meat, a small amount of oxygen is still present in the muscle cells at the early stage of postmortem muscle, enabling glucose to undergo oxidative metabolism. When all of the oxygen is consumed, the metabolic pathway switches to anaerobic metabolism to generate ATP for cellular activities. Lactic acid is produced and accumulated, resulting in a pH drop. However, due to drastically lowered expression of the PDK4 gene, the initial conversion of pyruvate to acetyl CoA and rate of oxygen consumption may be greater in PSE meat. The rate of metabolic switch from aerobic to anaerobic may be faster, resulting in a rapid pH drop that causes protein denaturation in the defective turkey meat. This hypothesis is supported by a lower pH at 15 min postmortem in PSE turkey (Chiang et al., 2008). Eadmusik et al. (2011) used 20-min postmortem pH to classify turkey breast samples as fast-glycolysing or normal-glycolysing muscle. Samples from the rapid glycolysing group had reduced solubility, which is an indicator of protein denaturation, but the ultimate pH was not significantly different from that of samples with normal rate of glycolysis. Their results support the hypothesis of the current study that although the magnitude of ultimate pH may influence meat quality as previously indicated (Barbut, 1993; Fernandez et al., 2002), a rapid rate of postmortem pH decline, due to decreased expression of PDK4, combined with high early-postmortem carcass temperature, may have a greater impact on protein denaturation associated with development of PSE turkey. ! 38 Differential expression of PDK4 has been observed in previous studies on meat quality. Lan et al. (2009) reported reduced mRNA abundance of PDK4 in muscle from commercial lean-bred Yorkshire pigs compared with a Chinese traditional breed. The authors claimed that the traditional breed has superior meat quality compared to the Yorkshire breed, but they did not directly measure meat quality indices between the two breeds. They suggested that PDK4 expression is correlated with fiber type; i.e., increased levels of PDK4 would be associated with a higher percentage of slow-twitch fibers. In this study, up-regulation of several genes in PSE turkey has been observed that would suggest a shift in expression from fasttwitch muscle proteins to slow-twitch muscle proteins (Tables 3.3 and 3.4). Slow twitch muscle generally has a greater degree of oxidative metabolism (Donoghue et al., 2007). The results are further consistent with this transition in that the mitochondrial genome was up-regulated in the defective meat (Table 2.3). An increased expression of PDK4 in association with increased expression of oxidativemetabolism genes has been expected to be found. However, the fact that PDK4 was down-regulated in PSE muscle suggests that in this subset of birds, there is impaired regulation of expression of this gene. In a study on differential expression in chickens differing in glycogen content, Sibut et al. (2011) observed down-regulation of PDK4 in high-glycogen chicken skeletal muscle compared with low-glycogen samples. They did not find differences in direct measures of meat quality such as drip loss, but the high-glycogen meat showed lower ultimate pH and lighter color (Sibut et al., 2011), determinants often been used as PSE indicators in previous studies (Oda et al., 2009; Ziober et al., ! 39 2010). However, although it is of interest that differential PDK4 expression was observed in the study of Sibut et al. their results are not directly comparable to the current study. They classified chicken muscle samples based on glycogen content to determine the molecular mechanisms involved in variation of meat quality. In this study, however, samples were first classified as PSE or normal and then analyzed for differential gene expression. 2.5 Conclusion In conclusion, this study is the first evidence of global differential gene expression between normal and PSE turkey. Pathway analysis shows several molecular signaling pathways associated with development of the turkey meat defect including the calcium signaling pathway which supports an abnormality of Ca 2+ homeostasis in the susceptible animals. The results also suggest skeletal muscle fast-to-slow isoform switch, which may reflect different molecular properties and disorganization of muscle fibers in PSE turkey. Altered stability of non-muscle actin polymer and actomyosin assembly in PSE turkey is implicated due to differential expression of genes in the RhoA signaling and actin cytoskeleton signaling pathway. Clearly, development of this turkey meat defect is complex and associated with interactions of more than one pathway. Here, evidence was found of dramatically lower expression of the PDK4 gene in PSE turkey meat, which may alter oxidative metabolism and be associated with an unusually high rate of postmortem metabolism in PSE turkey. Further investigation at the protein level and activity of this enzyme is needed to better understand its role in turkey PSE meat. ! 40 CHAPTER 3 DEEP TRANSCRIPTOME SEQUENCING REVEALS DIFFERENCES IN GLOBAL GENE EXPRESSION BETWEEN NORMAL AND PALE, SOFT AND EXUDATIVE TURKEY MEAT ABSTRACT An association between development of pale, soft and exudative (PSE) turkey and changes of gene expression has been shown; however, understanding of altered molecular mechanisms associated with the differentially expressed genes remains unclear. The objective of this study was to utilize deep transcriptome RNA sequence analysis (RNA-Seq) to identify differentially expressed genes and the associated molecular mechanisms between normal and PSE turkey breasts. Following collection of turkey breasts (n = 43), meat quality characteristics were determined and samples were classified as normal or PSE based on marinade uptake (high = normal; low = PSE). Total RNA from breast muscle samples with the highest (n=4) and lowest (n=4) marinade uptake were isolated and sequenced using the Illumina GAIIX platform. Of 21,340 gene loci discovered by RNA-Seq, 8480 loci completely matched the turkey reference genome, and 494 genes were differentially expressed (false discovery rate, FDR<0.05). Changes in gene expression were confirmed using quantitative real-time PCR. Pathway analysis of differentially expressed genes suggested abnormalities of calcium homeostasis and signaling pathways regulating actin cytoskeleton structure as well as carbohydrate metabolism and energy production in PSE samples. A dramatically decreased expression of pyruvate dehydrogenase kinase, isozyme 4 may alter early-stage postmortem glucose oxidative metabolism and result in development of PSE turkey meat. ! 41 3.1 Background Over the past 60 years, pale, soft and exudative (PSE) meat has been one of the greatest challenges to the poultry meat processing industry. The incidence of PSE turkey has increased in concert with intensive selection of turkeys for rapid growth rate and large breast size (Dransfield and Sosnicki, 1999; Owens et al., 2000; Samuel et al., 2012). Depending on environmental conditions, the incidence can range from < 1% to as high as 50% in commercial processing plants (Woelfel et al., 2002). The term PSE has been used to describe characteristics of defective meat with unusually light color and inferior water binding properties (Aberle et al., 2001). Myofibrillar proteins of PSE meat undergo irreversible denaturation and lose their most critical functionalities, including protein solubility, extractability and water holding capacity (Pietrzak et al., 1997; Warner et al., 1997). The loss of protein functionality is most readily manifested in processed meat products by excessive purge and structural defects. Therefore, the PSE problem results in substantially reduced processing yields and economic losses in the poultry industry (Owens et al., 2009). It is widely accepted that the extensive protein denaturation of PSE meat is a consequence of muscle hypermetabolism during early-stage postmortem conversion of muscle to meat. As a result of aberrantly rapid glycolysis, lactic acid is produced which accumulates inside the muscle cell at an accelerated rate, resulting in fast pH ! 42 decline. The combination of high carcass temperature and high acidity during the early-stage postmortem period causes protein denaturation (Barbut et al., 2008). In PSE pork, accelerated postmortem metabolism is associated with an abnormality of the sarcoplasmic reticulum (SR) Ca 2+ release channel protein, ryanodine receptor 1 (RYR1) (Mickelson et al., 1988b). Later, Otsu et al. (1994) showed that a single point mutation of RYR1 is responsible for a leak of the SR Ca 2+ release proteins. In turkeys, Chiang et al. (2007) identified alternatively spliced products of !RYR (homologous to mammalian RYR isoform 1) and single-nucleotide polymorphisms of the "RYR isoform (homologous to mammalian RYR isoform 3), 2+ suggesting that there may be differences in regulation of intracellular Ca . A delay in up-regulation of !RYR, "RYR and calsequestrin was also found in PSE turkey meat from birds subjected to heat stress (Sporer et al., 2012). Recently, global gene expression in PSE and normal turkey meat was profiled using the 6K turkey skeletal muscle long oligonucleotide (TSKMLO) microarray (Malila et al., 2013). The results suggested an association between differential 2+ expression of genes involved in Ca homeostasis with development of PSE turkey. Moreover, microarray analysis revealed a complex manifestation of several affected cellular pathways in addition to calcium signaling that were associated with PSE development in turkey. However, the microarray approach was limited by the fact ! 43 that some genes and alternatively spliced products, which are not included on the array, may significantly contribute to the development of turkey PSE. Deep transcriptome mRNA sequence analysis (RNA-Seq) is an emerging technology for transcriptome profiling and offers advantages over microarrays. Unlike hybridization-based microarray, RNA-Seq can quantify the actual gene expression (Wang et al., 2009) and identify transcripts beyond those represented on the microarray (Roh et al., 2010). Thus, RNA-Seq can reveal novel transcripts as well as novel isoforms produced by alternative splicing (Baginsky et al., 2010; Tranpnell et al., 2010). The hybridization-based microarray technique is also limited by dynamic signal range resulting from saturation of the fluorescence signal (Roberts et al., 2011). Additionally, results have shown that RNA-Seq data are more reproducible with smaller technical variation compared to microarray hybridization data (Marioni et al., 2008). The objective of this study was to utilize RNA-Seq to investigate global gene expression in normal and PSE turkey breast meat. Because RNA-Seq and microarray platforms offer distinct comparative advantages and disadvantages, this RNA-Seq study was designed to augment and complement the results obtained from the microarray study. The present study used RNA samples previously used for the microarray analysis from turkeys of the Randombred Control Line 2 (RBC2), representative of the late 1960s commercial turkeys (Nestor et al., 1967; Nestor, 1977a; Nestor, 1977b) and maintained without selection pressure at The Ohio Agricultural Research and Development Center of The Ohio State University ! 44 (Wooster, OH). Analysis of normal and PSE breast samples from the RBC2 line will serve as a basis for future comparisons with modern growth-selected turkey lines. 3.2 Materials and methods Sample information Turkey breast muscle samples used in this study comprised a subset of the control sample group collected from RBC2 turkeys in the study of Chiang et al. (2008). Briefly, the birds were slaughtered at 22wk of age using standard industry practices at the Michigan State University Meat Laboratory. Breast muscle was collected, snap frozen and stored at -80°C until utilized for RNA isolation or microsomal membrane protein preparation. Meat quality indices were also determined (Chiang et al., 2008). The samples were classified as normal or PSE based primarily on marinade uptake and secondarily on cook loss as previously described (Malila et al., 2013; Sporer et al., 2012). In brief, normal samples were classified based on high marinade uptake and on low cook loss. Conversely, PSE samples were characterized by low marinade uptake and high cook loss. Four samples with the highest marinade uptake (normal) and four samples with the lowest marinade uptake (PSE) were used (a subset of the samples evaluated by microarray; Malila et al., 2013). Meat quality indices of the samples utilized in this study are shown in Table 3.1. Total RNA samples were the same materials isolated from each muscle sample, and used for the microarray study (Malila et al., 2013). The RNA isolation procedure was detailed in Malila et al. (2013). Sample quality was assessed using ! 45 an Agilent Bioanalyzer and all samples had an RNA Integrity Number (RIN) above 8.0 (maximum = 10). Table 3.1 Meat quality indices of the meat samples utilized in the RNA-Seq study Meat quality index 1 pH (15 min postmortem) % Marinade uptake % Drip loss % Cook loss 2 Normal 5.77 ± 0.35 PSE 5.59 ± 0.19 Significance level 50.83 ± 9.64 0.46 ± 0.30 27.70 ± 2.18 15.19 ± 1.89 0.86 ± 0.35 31.16 ± 0.42 *** NS NS * 1 Meat quality was determined in the study of Chiang et al. (2008). The results shown here are specific for the meat samples defined as normal (n=4) or PSE (n=4) that were utilized in this study. 2 NS = non-significant (p>0.05), *P<0.05, **P<0.01, ***P<0.001 RNA transcriptome sequencing analysis RNA deep sequence analysis was performed at the Michigan State University Research Technology Support Facility (RTSF). Eight RNA libraries (meat quality levels = 2, biological replicates = 4) were constructed from 10 µg of total RNA per sample using the Illumina mRNA Sequencing Sample Preparation Kit (Illumina, Inc., San Diego, CA) according to the maunfacturer’s protocol. Briefly, mRNA was purified using oligo-dT beads (Illumina, Inc.) and subsequently fragmented. First strand cDNA was synthesized with random hexamer priming followed by second strand synthesis by random hexamer priming. Ends of the synthesized cDNA were repaired, A-tailed and bar-coded by adapter ligation. The cDNA templates were purified and enriched using PCR. All eight libraries were pooled and sequenced across 4 lanes. Paired-end reads were generated using the Illumina Genome ! 46 AnalyzerIIx (Illumina, Inc) with 2x55 bp. The resulting mRNA-Seq sequencing reads were aligned to the domestic turkey reference genome (Turkey_2.01, UMD, Ensemble release 61; Dalloul et al., 2010) using TopHat (Trapnell et al., 2009). Alignments produced from each library were used as input to Cufflinks (Trapnell et al., 2010), which assigned reads to isoforms and generated a set of predicted transcripts based on assembly of overlapping reads, which were subsequently transferred to Cuffcompare to be combined into a single, non-redundant set of transcript models. The resulting alignments were compared to an annotation file and the number of fragments uniquely aligned to each gene in the annotation were reported using HTseq-count (http://www-huber.embl.de/users/anders/HTSeq/). Read counts per locus generated by HTseq-count were used as input for Bioconductor R package, DESeq (Anders and Huber, 2010) which subsequently calculated sample variance from biological replicates and tested for differential expression following the standard method outlined in the DESeq documentation. Differential expression was reported as fold change (FC) indicating ratio of expression of the gene in PSE samples relative to normal samples. Negative FC denotes down-regulation while positive FC suggests up-regulation. RNA-Seq data have been submitted to the National Center for Biotechnology Information’s Gene Expression Omnibus (NCBI GEO) with GEO accession number GSE37176. Confirmation of expression patterns Fourteen genes, differentially expressed (false discovery rate; FDR<0.05) between normal and PSE samples, were selected for expression confirmation using ! 47 quantitative real-time PCR (qPCR). Of 14 genes, nine genes were representatives of genes with large fold changes (|FC| > 3.5). The other five were randomly chosen from the group of differentially expressed genes with moderate to low fold change (1.5 < |FC| < 3.0). Primers (Table 3.2) were designed using Primer Express 3.0 software (Applied Biosystems). Primer specificity was confirmed by subjecting the primer sequence to an NCBI BLAST search. Only primers that specifically matched their respective genes were submitted for primer synthesis by Operon Inc. (Huntsville, AL). Additional specificity was verified by dissociation curves with a single peak at melting temperature and no peak in no-template controls. The qPCR protocol was previously described in Malila et al. (2013). Functional and pathway analysis Pathway analysis was performed to identify biological functions of differentially expressed genes between PSE and normal samples. Genes with FDR<0.05 were uploaded into Ingenuity Pathway Analysis (IPA) software (http://www.ingenuity.com, Ingenuity Systems, Redwood City, CA). Of 494 differentially expressed genes, 402 were recognized by IPA and included in the analysis. Lists of functional categories, canonical pathways, and networks associated with differentially expressed genes were generated. ! 48 Table 3.2 Primer information for genes chosen for confirmation of expression identified in the RNA-Seq analysis using qPCR Gene ID Primer Length (bp) ATP1B4 F: TACCCTGGAAACGGCACATT R: TGTAGTTGACGTGCGTGAGCTT ACTB F: GTCCACCTTCCAGCAGATGTG R: CAATGGAGGGTCCGGATTC CCDC135 F: ATTACCTGGCACCTTTTCTTATTCA R: CGGAGGGCCTGCCTTTT COL6A1 F: TTCCATTGGTGCTCTTGCTATG R: TTTGGGATGATGGCGATACC DCN F: CCCATCGGATGAATATGTGAAA R: GCAATGTTCAACTAGGCCTTACTG FMOD F: GGCTCTCTACCAACACTTTCAACA R: GGTTCCCTTGAAGGTAGAGGTTCT FSCN1 F: CCGCTCCTCCTATGATGTCTTC R: CCCAGTGGTATCCTTGATGTTGT FSTL1 F: TGTCCTCCAGTGCTGATCTCAA R: ATGTAAATATCTGTGACTGGGCAACT NOV F: AATGATGGGCGGTGCTGTA R: GACAGCGGAACTCAACTTGAATG PDE10A F: GGGAGAAAGTAACGAGAGGTGAAG R: GAGGTTCCAGGAGCAATGGA PDK4 F: ATGAATGTCTGTAATAGTGCTTGCAA R:CATGTCTTCATTGTATGTTCTGCATATAC PFKFB3 F: CCAACTAAAGAGAACAATCCAAACAG R: GGGCCTTCCACTGCTCATAG POSTN F: CAGGGAGCTGGAACTGAGTACA R: TACAGGTGCTTCTTCCAAATGGA RGS2 F: AGGCTCCCAAGGAGATAAACATT R: TCCTGGAGGTTCTGTGCTATCA TAGLN F: CCCTCACCGATTGGTATTTATTG R: TTGTTGGTCCCCATCTGTAAGC ! 49 Tm (°C) 70 80 71 79 67 79 79 78 96 74 120 83 70 79 68 78 61 80 70 79 90 74 65 82 105 77 64 79 78 82 Microsomal membrane preparation Meat samples utilized for microsomal membrane preparation were from the same RBC2 turkey breast muscles previously identified as normal (n = 4) and PSE (n = 4) and used in RNA-Seq analysis. For the membrane preparation, 6 g breast muscle, excised immediately at slaughter and frozen at 80°C, was incubated in 30 mL of Buffer I (0.1 M NaCl, 5 mM Tris-maleate (pH 7.0), 30 mM !-mercaptoethanol) supplemented with protease inhibitor cocktail (0.2 mM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride, 1 µg/mL aprotinin, 1mM benzamidine, 1 µg/mL leupeptin, 1 µg/mL pepstatin A) for 15 min and homogenized using a Waring blender (15 sec on followed by 15 sec rest, 8 cycles). All steps of microsomal membrane preparation were carried out at 4°C. The homogenate was centrifuged at 3,300xg for 30 min and the supernatant was filtered through glass wool and centrifuged at 16,000xg for 30 min. The resulting pellet was re-suspended in 30 mL of Buffer II (0.6 M KCl, 5 mM Tris-maleate (pH 7.0), 30 mM !-mercaptoethanol) supplemented with the protease inhibitor cocktail described above, and gently homogenized using a Dounce homogenizer. The homogenate was centrifuged at 130,000xg for 40 min and the pellet collected and re-suspended in 0.5 mL 10% (w/v) sucrose. Concentration of crude microsomal protein was determined using the Pierce Bicinchoninic acid (BCA) Protein Assay kit (Thermo Scientific, Rockford, Illinois) in accordance with manufacturer’s recommendation. ! 50 Protein electrophoresis Crude microsomal protein was separated by electrophoresis. Briefly, 40 µg crude microsomal protein was mixed with 3.4 µL of 6X Laemmli sample buffer (BioRad, Hercules, CA). The protein solution volume was adjusted to 20 µL with water and heated at 90°C for 10 min. Samples were separated using a 12% Tris-glycine Criterion gel with dimensions of 133 x 87 x 1 mm (Bio-Rad) at constant 200 V for 1 hr at 4°C. The running buffer was Tris/glycine/sodium dodecyl sulfate (SDS) containing 25 mM Tris, 192 mM glycine, and 0.1% SDS, pH 8.3. Pre-stained lowrange molecular weight protein standards (Bio-Rad) were used as protein markers. Immunoblot After electrophoresis, the gel was equilibrated in 100 mL transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, pH 8.3) for 10 min at room temperature. Proteins were transferred to a methanol-prewetted Immobilon-FL polyvinylidene difluoride (PVDF) membrane (Millipore, Burlington, MA) using Criterion blotting cell (Bio-Rad) at constant 30 V for 4 h at 4°C. After blotting, the membrane was briefly rinsed with Tris-saline-tween buffer (TBST; 20 mM Tris, 140 mM NaCl 0.1% tween 20, pH 7.4) for 5 min at room temperature, and pre-incubated with 20 mL blocking buffer (LI-COR, Lincoln, NE) for 1 h at room temperature. The membrane was subsequently incubated overnight at 4°C with a mixture of a polyclonal rabbit antiPDK4 (101-114) antibody (Sigma-Aldrich, St. Louis, MO) mixed with a monoclonal mouse ! anti-!-actin antibody (Sigma-Aldrich) 51 diluted (1:500 and 1:10000, respectively) in the LI-COR blocking buffer. The blot was washed using TBST (5 min, room temperature, 4 times) and then incubated with a mixture of secondary antibodies (1:3000 anti-rabbit IgG IRDye 700DX® conjugated antibody, 1:3000 antimouse IgG IRDye800® conjugated antibody, Rockland, Gilbertsville, PA) diluted in blocking buffer in the dark for 1 h at room temperature. The membrane was scanned using the Odyssey® Infrared Imaging System (LI-COR). Images were acquired using Odyssey 3.0 analytical software (LI-COR). For analysis of the immunoblot images, densitometry was performed using ImageJ 1.46r (Rasband, 2012). A rectangular box was manually placed around the band of interest in the first lane. Features with identical size were used to measure the other bands for all lanes. The band intensity of PDK4 was normalized with the same-lane intensity value of !-actin, and the ratios were defined as relative protein abundance. Previous studies in our laboratory have verified !-actin as a valid endogenous control for PSE and normal turkey breast muscle (Unpublished data). The assay was conducted in triplicate; thus, there were three ratios, obtained from different blots, per biological replicate. An average of the three ratios was used as protein abundance of each bird in statistical analysis. Statistical difference in expression of PDK4 between normal and PSE turkey was determined using Student’s t test. ! 52 3.3 Results Differential gene expression between normal and PSE turkey skeletal muscle revealed by RNA-Seq In the current RNA-Seq analysis of turkey breast meat, Cuffcompare assembled mRNA sequence reads into 21,340 transcript loci and assigned Cufflinks class code to each locus (Table 3.3). The class code guided how the Cuffcomparegenerated locus matched to the reference gene. The top three class codes were st assigned to loci that completely matched with the intron chain (1 nd or were contained in the reference gene (2 class code, “=”), class code, “c”), or where at least one rd splice junction of the locus was shared with a reference transcript (3 class code, “j”). Of 21,340 loci, 8,480 loci were assigned to the first three classes, suggesting a close match of the loci with the reference gene. As a result, this group was subsequently utilized in identification of differential gene expression (Table 3.4). ! 53 Table 3.3 Number of loci classified within different Cufflink class code Cufflink Code class code priority Number of loci Code definition = 1 locus completely matched with intron chain 3346 (16%) c 2 locus contained in reference gene 3225 (15%) j 3 locus is potentially novel isoform and at least on splice junction is shared with a reference transcript 1909 (9%) e 4 a possible pre-mRNA fragment i 5 a transfrag falling within an intron region o 6 generic overlap with reference 357 (2%) p 7 possible polymerase run-on fragment 1637 (8%) x 8 s 9 . 10 216 (1%) exonic overlap with opposite strand of the reference intronic overlap with opposite strand of the reference likely due to mapping error loci with multiple classifications 2139 (10%) 2745 (13%) 451 (2%) 5315 (25%) Total 21,340 Table 3.4 Number of differentially expressed transcripts at different significance level Significance level FDR < 0.01 0.01 " FDR < 0.05 326 (3.84%) 168 (1.98%) 0.05 " FDR < 0.1 0.1 " FDR < 0.2 0.2 " FDR < 0.3 0.3 " FDR < 0.4 0.4 " FDR < 0.5 FDR # 0.5 133 (1.57%) 210 (2.48%) 198 (2.33%) 229 (2.70%) 397 (4.68%) 6,820 (80%) Total ! Number of differentially expressed transcripts 8,481 54 Differential expression of 494 loci between normal and PSE turkey skeletal muscle was revealed (FDR<0.05), with 182 down-regulated loci and 312 upregulated loci (Appendix B). Approximately 80% of the down-regulated loci fall within FC range between -1.5 to -3.0, and the remaining 20% showed FC < -3.0. About 86% of up-regulated loci showed FC within the range of 1.5 to 3.0 with the remaining 14% of up-regulated loci showing FC > 3.0 (Figure 3.1). Among differentially expressed loci, 91 are unknown or uncharacterized proteins. For the identified loci, pyruvate dehydrogenase kinase isoform 4 (PDK4) showed the greatest downregulation (FC = -14.1) in PSE turkey while nephroblastoma overexpressed gene (NOV) was the most up-regulated gene (FC = 37.7) in PSE turkey meat. Biological functions and location of the changed genes showing the greatest differential expression are shown in Figure 3.2. ! 55 Figure 3.1. Distribution (%) of differentially expressed transcripts (FDR<0.05) at different fold change (FC) ranges. Percentage was calculated separately within down-regulated (total = 182) and up-regulated (total = 312) transcripts. Different shades on bar graphs represent ranges of absolute value of FC (|FC|). ! 56 Figure 3.2 Schematic diagram showing localization and associated functions of the most up-regulated and down-regulated genes identified by RNA-Seq. The most up-regulated gene is nephroblastoma overexpressed gene (NOV). The most down-regulated gene is pyruvate dehydrogenase kinase isozyme 4 (PDK4). FC, fold change or gene expression in PSE relative to normal samples. Negative FC indicates down regulation of gene in PSE samples; positive indicates up-regulation in PSE samples. ! 57 Confirmation of gene expression differences by qPCR technique Differential gene expression, observed by RNA-Seq, was confirmed by qPCR analysis for fourteen genes (Figure 3.3). Overall, the fold changes of most genes analyzed by qPCR were similar to that observed by RNA-Seq at the same significance level ($ = 0.05). Only fascin (FSCN), follistatin-related protein 1 (FSTL1), fibromodulin (FMOD) and NOV showed expression differences between PSE and normal meat (FDR < 0.05) by RNA-Seq but were not significantly different by qPCR analysis. This discrepancy could have resulted from biological variation. Additionally, because of the principles underlying each technique, it is possible that RNA-Seq and qPCR were actually reporting expression of different isoforms but those isoforms were still annotated as the same gene. ! 58 Figure 3.3 Confirmation of gene expression analyzed by RNA-Seq using qPCR. Results are presented as relative expression or fold change for gene expression in PSE relative to normal samples. Bars below the origin indicate lower expression (down-regulation) of the gene in PSE samples; bars above the origin indicate higher expression (upregulation) in PSE samples. Statistical significance indicates change in expression between PSE and normal samples within each technique (*FDR < 0.05 for RNA-Seq, † P < 0.05 for qPCR). ! 59 Functional and pathway analysis For pathway analysis, 402 genes identified by RNA-Seq and recognized by IPA, were grouped on the basis of biological and functional relationships into canonical pathways, functional networks and cellular and molecular functions. The current RNA-Seq study confirmed abnormality of calcium signaling pathways (Figure 3.4) associated with development of PSE turkey. Actin cytoskeleton and cell motility-related canonical pathways exhibited a greater degree of complexity of interactions compared to that reported in the microarray study (Malila et al., 2013). The interactions comprise actin cytoskeleton signaling and Ras homology family member A (RhoA) signaling, two pathways originally identified in the microarray analysis (Malila et al., 2013), which now include integrin, integrinlinked kinase (ILK), focal adhesion kinase (FAK), tight junction, epithelial adherens junction and remodeling of epithelial adherens junction signaling pathways (Figure 3.5). The top three altered biological functions associated with differential gene expression in PSE turkey were cellular movement, cellular assembly and organization, and cellular function and maintenance. These three functional activities potentially increased in the defective meat (Table 3.5), as predicted by IPA based on overall directional changes in expression of genes associated with particular function. Ingenuity Pathway Analysis also predicted likelihood of increased metabolism of carbohydrate and fatty acids in PSE samples. ! 60 Gene ID MYH13 FC -5.0* MYH10 1.5* MYL2 1.6* CREB5 1.8* ACTA2 1.9* MYL9 1.9* MYH11 2.4* CHRND 3.4* ACTC1 7.2* Figure 3.4 Diagram of the calcium signaling pathway associated with development of PSE turkey. The pathway, suggested by the IPA, showed 2+ interactions among differentially expressed genes associated with regulation of Ca concentration between normal and PSE turkey skeletal muscle. FC, fold change or gene expression in PSE relative to normal samples. Negative FC indicates down regulation of gene in PSE samples; positive indicates up-regulation in PSE samples. *FDR<0.05. ! 61 Figure 3.5 Diagram indicating overlapping canonical pathways that regulate actin cytoskeleton and cell motility, and their associated differentially expressed genes (FDR<0.05). Genes in bold show down-regulation in PSE turkey. (A) Integrin-mediated signaling pathways modulated cellular activities and signal transduction upon interactions of integrin, integrin-linked kinase (ILK), focal adhesion kinase (FAK), Ras homology family member A (RhoA) and actin cytoskeleton signaling pathways; (B) Tight junction signaling pathway interacted with RhoA and actin cytoskeleton signaling pathways and modulates tight junction complex; and (C) Epithelial-associated signaling pathways linked to actin cytoskeleton signaling, leading to actin reorganization and adheren junctions formation. Together, alteration of this signaling pathway network affects biological functions in PSE turkey. ! 62! Table 3.5 Potential downstream activities of the top altered biological functions associated with differential gene expression in PSE turkey. Biological function Cellular movement Potential 1 activation Activation 2 z-score Number of molecules Increased 3.34 99 Increased Increased 2.20 1.98 8 25 Increased 1.76 44 Increased 2.38 37 Increased Increased 1.76 1.60 28 28 Cellular assembly and organization - Rearrangement of cytoskeleton - Growth of plasma membrane projections - Microtubule dynamics Cellular Function and maintenance - Formation of cellular protrusions Energy production - Carbohydrate metabolism - Fatty acid metabolism 1 An increase or decrease in a biological process was predicted by Ingenuity Pathway Analysis (IPA) based on experimentally derived relationships between genes or proteins and directional change in expression of genes in the dataset. 2 Activation z-score was calculated by IPA. The greater z-score implies the higher possibility of increasing activity. ! 63! Immunoblot of PDK4 protein Upon identification of differentially expressed genes, PDK4 was clearly of particular interest due to its substantial down-regulation in PSE samples observed in both RNA-Seq and the TSKMLO microarray analyses (Malila et al., 2013). The association of decreased PDK4 expression and altered glucose oxidation in PSE turkey was proposed by Malila et al. (2013). Additionally, pathway analysis of the current RNA-Seq analysis emphasized aberrant energy metabolism arising from differential gene expression in PSE turkey based on the following observations. First, energy production was highlighted as the first network identified by IPA. Second, based on the direction of differential gene expression, IPA predicted an increase in carbohydrate and lipid metabolic activities in PSE samples (Table 3.5), which is consistent with the hypothesis of hypermetabolism in the susceptible turkeys. Taken together, PDK4 was prioritized in this study for further investigation at the translational level. Abundance of PDK4 protein in turkey breast meat was determined using protein immunoblot assay (Figure 3.6). The results demonstrated that PDK4 protein was 3.4 fold lower in PSE meat compared to normal samples (P < 0.001). ! 64! Figure 3.6 Expression of PDK4 protein in turkey breast muscle. (A) Protein immunoblot shows PDK4 protein abundance in normal (n = 4) and PSE (n = 4) turkey meat. Each lane presents analysis of proteins collected from an individual bird. Beta-actin was used as an endogenous control. Blot is a representative from three technical replicates. (B) Histograms, constructed from densitometry of immunoblot results, represent difference in abundance of PDK4 protein expressed in normal and PSE turkey. Expression of PDK4 protein is presented as ratio of band intensity between PDK4 and !-actin within lane. Average expression of PDK4 was calculated from results of three different blots. ***P < 0.001 ! 65! 3.4 Discussion Previous studies of differential gene expression regarding development of 2+ PSE turkey were mainly focused on individual SR Ca regulators (Chiang et al., 2004; Chiang et al., 2007; Oda et al., 2009; Sporer et al., 2012). However, the results suggested that these single-gene approaches were inadequate in addressing the evident complex etiology of this meat defect. Recently, a microarray specifically developed to analyze gene expression in turkey skeletal muscle was used to profile the transcriptomes of PSE and normal turkey breast meat samples (Malila et al., 2013). Differential expression of many genes and interactions among molecular signaling pathways were observed which offered new insights into the mechanism of PSE meat development. In the current study, the Illumina GAIIX RNA-Seq platform was utilized to complement and extend our previous microarray study. The RNA-Seq analysis revealed 494 differentially expressed transcripts (FDR < 0.05) and multiple associated signaling pathways associated with development of PSE turkey. Despite the inherent advantages of RNA-Seq over microarray, we also experienced some challenges in RNA-Seq data analysis. Some differentially expressed transcripts identified in the microarray study (Malila et al., 2013) were classified in this RNA-Seq analysis with a low priority Cufflinks class code, or were not present in the current turkey genome assembly. Because the turkey reference genome is still an early draft, these results are not surprising. In addition, some large gene families such as myosin comprise closely related individual genes that are not yet well annotated in ! 66! the turkey reference genome; thus, loci generated from the RNA-Seq read alignments against these reference genes are not well defined. In contrast, probes on the TSKMLO microarray were designed based on turkey skeletal muscle cDNA libraries constructed and annotated with the chicken genome because the turkey reference genome was not available (Reed et al., 2008). Since the chicken genome is more established, the TSKMLO microarray study may identify genes that are not yet annotated in the turkey reference genome (Malila et al., 2013). However, the RNA-Seq data can be realigned with updated versions of the turkey reference genome as new versions become available in the future. The current RNA-Seq analysis confirmed the important roles of actin cytoskeleton signaling, RhoA signaling, and the calcium signaling pathways associated with development of PSE turkey. The findings extended interactions within and between actin cytoskeleton and RhoA pathways to a larger network of integrin-related signaling pathways (Figure 3.5), supporting the role of irregular actin cytoskeletal filaments in PSE turkey. The turkey PSE problem is clearly multifactorial with respect to the involvement of various genes and molecular pathways. The hypothesis of fast-to-slow skeletal muscle isoform conversion in PSE turkey, originally proposed in the microarray study, was supported in this study by downregulation of fast-switch myosin heavy chain (MYH; MYH13, FC = -5.0), and upregulation of slow-switch isoforms (MYH10, FC = 1.5; MYH11, FC = 2.4) in PSE samples. For the calcium signaling pathway, a greater mRNA abundance of the nicotinic cholinergic receptor gene (CHRND) was observed in PSE turkey (FC = ! 67! 3.4). This gene encodes the delta subunit of the muscle acetylcholine receptor (AchR). In skeletal muscle, the acetylcholine receptor, a pentameric protein composed of five subunits with the stoichiometry "2!#$, is located at the neuromuscular junction (Goldman et al., 1988). In neuromuscular signal transduction, acetylcholine binds to the receptor, which then undergoes a conformational change, leading to opening of an ion-conducting channel across the plasma membrane (Pedersen and Cohan, 1990). A mutation of CHRND has been implicated in functional disruption of acetylcholine receptor, leading to an abnormality of muscle contraction (Engel and Sine, 2005). Activity of this receptor requires coordination of all subunits; thus, the increased expression of CHRND in PSE samples may affect the function of acetylcholine receptor and alter ionic 2+ homeostasis, including Ca , in the skeletal muscle cell. This is the first evidence indicating that aberrant Ca membrane Ca 2+ 2+ regulation in PSE turkey may involve a plasma channel. Malila et al. (2013) reported a substantial down-regulation of PDK4 in PSE turkey. This gene encodes an enzyme that catalyzes phosphorylation of the alpha subunit of mitochondrial pyruvate dehydrogenase complex (PDH) into an inactive form, thus inhibiting conversion of pyruvate into acetyl CoA (Popov, 1997; Wynn et al., 2008). Based on the previous findings, we hypothesized that down-regulation of PDK4 may result in a decreased level of PDK4 enzyme in the susceptible animals. Therefore, at the early-stage of postmortem muscle metabolism, the flux of conversion of pyruvate into acetyl CoA by PDH would be greater in PSE samples. ! 68! Oxygen depletion may be faster. Thus, the metabolism switches from oxidative to anaerobic faster in the susceptible animals, leading to a rapid pH drop (Figure 3.7). High acidity combined with high carcass temperature at the early postmortem-stage is associated with protein denaturation in the defective turkey meat. Although a large fold-change of PDK4 was observed in the microarray study (Malila et al., 2013), the change was not statistically significant. Herein, RNA-Seq revealed differential expression of PDK4 (FC = -14.1) at FDR < 0.05. Immunoblot assays of PDK4 protein (Figure 3.6) also demonstrated low abundance of PDK4 in PSE samples (P < 0.001). This result strongly supports the hypothesis of dysregulation of postmortem oxidative glucose metabolic pathways. ! 69! Figure 3.7 Schematic diagram representing mitochondrial oxidative metabolism altered by decreased expression of PDK4 in PSE turkey. Mitochondrial pyruvate dehydrogenase complex (PDH) is regulated by pyruvate dehydrogenase phosphatase (PDP) or PDK4, thus contributing to the regulation of glucose metabolism. Down-regulation of PDK4 gene expression in PSE turkey may alter regulation of conversion of pyruvate to acetyl CoA and accelerate early stage postmortem oxidative metabolism. In the current study, pathway analysis revealed altered metabolic processes of not only carbohydrate, but also lipid. Based on the role of PDK4 in regulating energy production (Figure 3.3), PDK4 may cross-communicate between glucose and fatty acid metabolisms. Interactions between glucose and fatty acid metabolisms in ! 70! muscle and adipose tissue have been proposed by Randle et al. (1963) as a glucose fatty-acid regulatory cycle. To maintain glucose level during starvation, glucose oxidation is switched off through phosphorylation of PDH by PDK4, and fatty acids are used as fuel in production of acetyl CoA and NADH (Randle et al., 1994). Sugden et al. (2001) also proposed that PDK4 acts as a mediator for fatty acid oxidation by increasing available pyruvate for oxaloacetate formation. In a study on differential gene expression in chicken breast muscle, Sibut et al. (2011) reported down-regulation of PDK4 in high-glycogen breast samples compared with lowglycogen samples. In contrast, when comparing high-fat traits relative to low-fat ones, they observed up-regulation of PDK4, suggesting a strong interaction between carbohydrate and lipid metabolism. However, it must be noted that in the study of Sibut et al., chicken muscle samples were classified based on glycogen content or fat content to determine the molecular mechanisms involved in variation of meat quality. In this study, the samples were first classified as PSE or normal, then analyzed and compared for differential gene expression. The change in expression of PDK4 may alter contraction and relaxation process of skeletal muscle. Recently, Herbst et al. (2012) observed a significant increase in initial force of PDK4-knock-out mouse skeletal muscle in response to 2+ moderate-intensity stimulation. This leads us to hypothesize that abnormal Ca regulation combined with down-regulation of PDK4 may provide an effect on extensive muscle contraction in PSE-susceptible birds that accelerates early-stage postmortem metabolism. ! 71! As in other transcriptome analyses, RNA-Seq depicts steady state transcript abundance which can be utilized to monitor changes in gene expression in response to biological stimuli. Such analyses reveal candidate gene sets that with further validation can subsequently be used for classification of samples. Although transcriptional level changes generally correspond to changes in translation, mRNA abundance does not always directly reflect expression of encoded proteins (Preiss et al., 2003). To gain a complete understanding of molecular mechanisms underlying development of PSE turkey, a series of studies must be completed for candidate genes at the protein level, including protein abundance and protein activity. In this study, we provided an example of a protein level investigation by performing a PDK4 immunoblot. Decreased PDK4 protein abundance in PSE meat corresponded to decreased transcript abundance. The results support the important role of PDK4 in development of PSE turkey. 3.5 Conclusions Study of global differential gene expression between normal and PSE turkey using RNA-Seq technique highlighted complex interactions of multiple cellular signaling pathways associated with differential gene expression in development of PSE turkey. The results from the current RNA-Seq complement and extend the previous findings from the microarray analysis. Difference in mRNA abundance of nicotinic acetylcholine receptor delta subunit suggests, for the first time, that aberrant calcium homeostasis may be based not only at the SR level, but also extracellular calcium regulation. Alteration of cellular signaling pathways associated with organization of actin cytoskeleton has been confirmed and the complexity of the ! 72! network has been extended. Substantial down-regulation of PDK4 at transcriptional and translational levels in PSE turkey supports the hypothesis of an abnormality in glucose oxidative pathway, which may result in an unusually high rate of postmortem metabolism in PSE turkey. Activities of PDK4 in PSE turkey must be further investigated for better understanding of its mechanisms underlying development of PSE turkey. The results also suggest PDK4 as a potential gene marker to identify susceptible turkeys from the population. ! 73! CHAPTER 4 EXPRESSION OF PYRUVATE DEHYDROGENASE KINASE ISOZYME 4 IN PALE, SOFT AND EXUDATIVE MEAT FROM RANDOMBRED AND MODERN COMMERCIAL TURKEYS ABSTRACT Fundamental mechanisms responsible for post-mortem hypermetabolism associated with development of PSE turkey remain unclear. Previous transcriptome analyses of turkey breast muscle from the randombred control line (RBC2), representative of turkeys from the 1960s maintained without selection pressure, revealed substantial down-regulation of the pyruvate dehydrogenase kinase isozyme 4 (PDK4) gene in PSE samples. PSE meat quality continues to be a problem in modern commercial (COMM) turkey lines genetically selected for increased breast muscle mass. We hypothesized that as in the RBC2 line, PDK4 is also downregulated in PSE turkey from COMM lines and that decreased PDK4 transcript levels causes decreased PDK4 protein abundance. The objective of this study was to determine transcript and protein levels of PDK4 in PSE and normal turkey muscle within both COMM and RBC2 lines. Turkey breast muscle samples were harvested from both lines (RBC2, n = 43; COMM n = 39), and meat quality characteristics were determined. Breast samples were classified as normal or PSE based on marinade uptake (high = normal; low = PSE). Turkey breasts from each line with the highest (n = 6) and lowest (n = 6) marinade uptake were utilized as normal and PSE samples. Change in expression of PDK4 was determined within line. Transcript abundance of PDK4 mRNA was quantified using quantitative real-time PCR. Significant downregulation of PDK4 was observed in PSE meat of both lines (P < 0.05) with a 4.3- ! 74! fold decrease in RBC2 and a 12.8-fold decrease in COMM turkey. Microsomal protein was prepared from the normal and PSE meat samples from each line and subsequently utilized for determination of PDK4 protein using immunoblotting assay. Decreases in PDK4 protein level were found with high significance (P < 0.001) in both lines (RBC2, FC = -3.4; COMM, FC = -2.6) and in agreement with transcript differences. The results support the biological relevance of PDK4 suppression in the development of PSE turkey meat, and also suggest that the mechanism responsible for the decreased PDK4 in RBC2 turkey subpopulations has been maintained in a commercial line. 4.1 Background Over the past several decades, the turkey breeding industry has worked to meet increasing consumer demand for poultry products by improving turkey growth rate and carcass muscle mass. Compared to the commercial turkeys of the 1960s, modern commercial turkeys are marketed in about half the time (reviewed by Barbut et al., 2008) and yield heavier breast muscle mass (Fernandez et al, 2001; Lilburn and Nestor, 1991; Updike et al., 2005). However, as a consequence of the successful turkey-breeding selection, the pale, soft and exudative (PSE) meat defect in turkey has become more prevalent (Dransfield and Sosnicki, 1999). Processed meat products made from PSE meat are typically inferior because of reduced protein functionality, including reduced solubility, reduced binding, and poor gelation properties. This problem poses an increasing challenge to processing plants as consumer preferences trend toward value-added, processed meat products instead of whole bird or bird cuts. ! 75! Development of PSE turkey is associated with an accelerated postmortem glycolysis that leads to rapid pH decline in the muscle cell (Pietrzak et al., 1997). It has been hypothesized that changes in such biochemical conditions of PSE turkey 2+ associated with intracellular [Ca ] overload due to a leak by sarcoplasmic reticulum calcium channel proteins that promotes hypermetabolism of skeletal muscle. Recently, transcriptome profiling studies of normal and PSE turkey breast meat collected from the randombred control line (RBC2), a non-selected line representing commercial turkeys of the 1960s (Nestor et al., 1967; Nestor, 1977a; Nestor, 1997b), indicated differential expression of genes involved in Ca 2+ regulation in muscle cells, supporting the hypothesis of an abnormality in calcium homeostasis in PSE turkey (Malila et al., 2013; Malila et al., in preparation). Additionally, the studies revealed changes in expression of multiple genes from several cellular signaling pathways, including energy production and actin cytoskeletal organization, as well as interactions among the associated pathways, indicating the complexity of development of PSE turkey. One finding of particular interest is down-regulation of the gene encoding pyruvate dehydrogenase kinase isozyme 4 (PDK4) identified in both microarray (fold change, FC = -25.9) (Malila et al., 2013) and deep transcriptome sequencing (RNASeq; FC = -14.1) platforms (Malila et al., in preparation). This enzyme inactivates the pyruvate dehydrogenase complex (PDH) through phosphorylation of serine residues located on E1" subunit of PDH, thus inhibiting conversion of pyruvate to acetyl CoA during glucose catabolism (Sugden and Holness, 2003). In concert with decreased ! 76! transcript abundance, there was a significant reduction of PDK4 protein abundance in PSE turkey from the RBC2 line (Chapter 3). Decreased expression of PDK4 at both transcriptional and translational levels in PSE meat from random-bred turkey suggests that an acceleration of early-stage postmortem metabolic rate in the susceptible birds may not only derive from accelerated glycolysis, but may also include an altered glucose oxidative metabolism. The determination of PDK4 expression was previously limited to a small subset of normal and PSE samples from RBC2 turkey. The objective of this study was to assess differences in PDK4 transcript and protein abundance within a larger RBC2 turkey sample size and to determine whether these differences persist within an intensively selected, modern commercial line. 4.2 MATERIALS AND METHODS Sample information Turkey breast meat was a subset of meat samples collected in the study of Chiang et al. (2008). In brief, RBC2 turkeys were obtained from The Ohio Agricultural Research and Development Center of The Ohio State University. Growth-selected COMM turkeys were obtained from a local turkey producer. The turkeys from RBC2 line (43 birds) were harvested at 22 weeks of age while COMM birds (39 birds) were harvested at 16 weeks of age (Chiang et al., 2008). Breast muscle samples were collected, immediately snap frozen in liquid nitrogen and stored at -80°C for RNA isolation and microsomal protein preparation. Meat quality ! 77! indices of the samples were determined in the study of Chiang et al. (2008). The samples were classified as normal or PSE primarily based on percent marinade uptake (high = normal, low = PSE) as previously described (Malila et al., 2013; Sporer et al., 2012). Six samples for each extreme of normal and PSE characteristics (n = 6) from each line were used in this study. RNA Isolation and quantitative real-time PCR (qPCR) Total RNA of RBC2 samples were the same materials isolated and utilized in the previous transcriptome analyses (Malila et al., 2013; Malila et al., in preparation). Total RNA was isolated from breast meat samples in accordance with the protocol previously described (Malila et al., 2013). Quantity of total RNA was measured using a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Integrity of total RNA was confirmed using an Agilent 2100 Bioanalyzer (Santa Clara, CA). Samples with an RNA Integrity Number (RIN) equal to or exceeding 8.0 (RIN = 10 is the best) were used in qPCR. Primers were designed using Primer Express 3.0 software (Applied Biosystems) and synthesized by Operon Inc. (Huntsville, AL). Primer sequences were the following: PDK4 (GenBank# ATGAATGTCTGTAATAGTGCTTGCAA-3’3472 NM_001199909) and forward reverse 5’33475’3536- CATGTCTTCATTGTATGTTCTGCATATAC-3’3508 with a product size of 90 bp; ACTB (!-actin, GenBank# AY942620) forward 5’1071-GTCCACCTTCCAGCAGATG TG-3’1091 and reverse 5’1132-CAATGGAGGGTCCGGATTC-3’1111 with a product ! 78! size of 71 bp. The qPCR protocol was described in Malila et al. (2013). Beta-actin was used as an endogenous control gene (Sporer et al., 2012). Expression of PDK4 in PSE samples relative to normal samples was calculated using 2 -%%CT method (Livak and Schmittgen, 2001). Microsomal membrane preparation Microsomal membrane samples were prepared as previously described (Malila et al., in preparation). All preparation steps were carried out at 4°C. Concentration of crude microsomal protein was determined using the Pierce Bicinchoninic acid (BCA) Protein Assay kit (Thermo Scientific, Rockford, Illinois) following manufacturer’s recommendation. Protein electrophoresis and immunoblot assay Electrophoresis and immunoblot analysis were conducted according to the procedure described in Malila et al. (in preparation). Briefly, crude microsomal proteins were separated by electrophoresis using a 12% Tris-glycine Criterion® gel with dimension of 133 x 87 x 1 mm (Bio-Rad, Hercules, CA) at constant 200 V for 1 hr at 4°C. The gel was subsequently equilibrated in 100 mL transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, pH 8.3) for 10 min at room temperature and blotted with methanol-prewetted Immobilon-FL polyvinylidene difluoride (PVDF) membrane (Millipore, Burlington, MA) using a Criterion blotting cell (Bio-Rad). The transfer of proteins was conducted at constant 30 V for 4 h at 4°C. After blotting, the ! 79! membrane was briefly washed with Tris-saline-tween buffer (TBST) for 5 min and blocked with 20 mL of the LI-COR blocking buffer (LI-COR, Lincoln, NE) for 1 h at room temperature. The membrane was then incubated overnight at 4°C with a mixture of primary antibodies (1:500 polyclonal rabbit Anti-PDK4 (101-114) antibody, Sigma-Aldrich, and 1:10,000 monoclonal mouse anti-!-actin antibody, SigmaAldrich). The blot was washed with TBST for 5 min at room temperature, repeating 4 times and then incubated with a mixture of secondary antibodies (1:3000 anti-rabbit IgG IRDye 700DX® conjugated antibody, 1:3000 anti-mouse IgG IRDye800® conjugated antibody, Rockland, Gilbertsville, PA) in dark for 1 h at room temperature. The membrane was scanned using the Odyssey® Infrared Imaging System (LI-COR). Image analysis Immunoblot image analysis was carried out as previously described in Chapter 3. In brief, densitometry was performed using ImageJ 1.46r (Rasband, 2012). A rectangular box was manually placed around the band of interest in the first lane. Features with identical size were used to measure the other bands for all lanes. Band intensity of PDK4 was normalized with the in-lane intensity value of !actin, and the ratio values was defined as protein abundance. B-actin was used as a control. ! 80! Because the assay was conducted in triplicates, there were three ratio values, obtained from different blots, per biological replicate. Average of the three ratio values was used as protein abundance of each bird in statistical analysis. Statistical analysis Statistical analysis in expression of PDK4 between normal and PSE turkey within genetic line was assessed using Student’s t test. Significant differences were identified at P < 0.05. 4.3 Results Meat quality indices of normal and PSE turkey meat Meat quality indices of normal and PSE turkey meat are shown in Table 4.1. The PSE meat from both lines exhibited significantly reduced marinade uptake and increased cook loss compared with normal samples (P < 0.05). Significant differences in 15 min postmortem pH and L* value were found only in COMM turkeys. Drip loss was not statistically different within either turkey genetic line. ! 81! Table 4.1 Meat quality indices of the meat samples utilized in the study of PDK4 expression in RBC2 and COMM turkey. Meat quality 1 index RBC2 2 COMM Normal PSE Normal pH15 5.72 ± 0.11 5.52 ± 0.08 6.02 ± 0.09 L* value % Marinade uptake % Drip loss 54.2 ± 0.65 56.2 ± 0.97 57.9 ± 0.54 45.3 ± 4.65 0.52 ± 0.10 % Cook loss 27.8 ± 0.84 1 2 ! a b b 17.2 ± 1.02 0.68 ± 0.16 a 30.7 ± 0.32 2 PSE a b a 25.9 ± 2.62 0.55 ± 0.35 b 28.2 ± 0.48 b 5.77 ± 0.05 a 61.0 ± 0.86 b 7.3 ± 1.30 1.44 ± 0.38 a 32.1 ± 0.62 Meat quality was determined in the study of Chiang et al. (2008). The results shown here are specific for the meat samples defined as normal (n = 6) or PSE (n = 6) that were utilized in this study. The results are presented as mean ± SEM. The pH was determined at 15 min postmortem. The other indices were determined at 24 h postmortem. Different superscript letters indicate significant difference (P < 0.05) between normal and PSE samples within genetic line. RBC2 = randombred control line, COMM = commercial line 82! Expression of PDK4 at transcriptional and translational levels in turkey meat Differences in mRNA abundance of PDK4 were observed in PSE turkey samples collected from both RBC2 and COMM lines (Figure 4.1). In RBC2 turkey, PDK4 was down-regulated by -4.3 fold (P < 0.05), whereas expression of PDK4 in PSE meat from COMM line was reduced -12.8 fold (P < 0.05). A decrease in PDK4 protein abundance in PSE turkey meat from both lines was observed in agreement with transcript difference (Figure 4.2). Reduced PDK4 protein showed high significance (P < 0.001) in both lines (RBC2, FC = -3.4; COMM, FC = -2.6). ! 83! Figure 4.1 Relative mRNA abundance of PDK4 between normal and PSE turkey as determined using qPCR. Bars represent mean %%Ct ± SEM of PSE relative to the normal sample within line. &-actin was used as an endogenous control. Expression of PDK4 decreased 4.3-fold in PSE samples of RBC2 line (n = 6) and 12.8-fold in the PSE samples of COMM line (n = 6). *P < 0.05. ! 84! (A) (B) Figure 4.2 Differences in abundance of PDK4 protein between normal and PSE turkey. (A) Immunoblots of PDK4 between normal (n=6) and PSE (n=6) turkey collected from RBC2 and COMM lines. (B) Histograms show mean±SEM calculated from band intensity of PDK4 normalized with intensity value of !-actin from the same lane. Statistical analysis was performed within genetic line. ***P < 0.001 ! 85! 4.4 DISCUSSION In response to high consumer demand for poultry products, turkey breeding programs have focused on enlarging muscle mass while enhancing growth rate. However, the success of the breeding program has coincided with the increasing prevalence of undesirable traits, including myopathies and skeletal deficiencies, which eventually have a detrimental effect on meat quality (Dransfield and Sosnicki, 1999). In this study, it is notable that marinade uptake of breast meat from modern growth-selected COMM turkeys was lower than that of random-bred RBC2 line regardless of normal or PSE meat classification. This evidence corresponds with previous reports indicating that turkey breeding selection program concentrating only on accelerated growth rate and muscling mass consequently worsens turkey meat quality (Dransfield and Sosnicki, 1999; Updike et al., 2005). The problem of PSE turkey was first documented in the 1970s (Vanderstoep and Richards, 1974; van Hoof, 1979). Development of this turkey meat defect is associated with postmortem hypermetabolic rate in the susceptible birds (Vanhoof, 1979; Pietrzak et al., 1997). In this study, no significant difference in pH at 15 min postmortem between normal and PSE turkey was observed in the RBC2 birds. However, this parameter of PSE samples from COMM line was significantly lower than that of normal samples (P < 0.05). The results suggest high rate of early postmortem metabolism in the susceptible birds from the modern line. ! 86! According to the previous global gene expression analyses (Malila et al., 2013; Malila et al., in preparation), substantial down-regulation of PDK4 was found in PSE meat from RBC2 line. Due to its roles in regulating glucose catabolism, this gene was of particular interest. We hypothesized that the lower PDK4 mRNA abundance in PSE meat may result in a decreased level of PDK4 enzyme. The change in amount of this enzyme may alter glucose oxidation and postmortem metabolic rate in the PSE turkey. In this study, the previous hypothesis is supported by the immunoblot assay showing that, in RBC2 samples, PDK4 protein was expressed 3.4-fold lower in PSE meat compared to normal samples (P < 0.05). The initial analysis of global gene differences between normal and PSE meat was conducted in RBC2 turkeys maintained without selection pressure, to serve as a basis for understanding the role of genetics of PSE development in modern growth-selected turkeys. In this study, expression of PDK4 was compared between normal and PSE turkey within genetic lines, RBC2 and COMM lines. For RBC2 samples, the sample size was expanded from the previous immunoblot analysis (Malila et al., in preparation) to cover all RBC2 samples utilized in the microarray study (Malila et al., 2013). Down-regulated expression of PDK4, at both transcriptional and translational levels, was clearly observed in PSE meat of both RBC2 and COMM line (P < 0.05). Reduced expression of PDK4 in both lines strongly confirms the biological relevance of PDK4 dysregulation in PSE turkey. The results also suggest that the decrease in ! 87! PDK4 expression has been maintained during breeding selection of the COMM turkeys, in which the problem of the PSE defect appears to be more pronounced. It can be speculated that, at the early postmortem stage, when oxygen is still present, glucose undergoes aerobic catabolism in order to generate ATP for cellular activities. However, due to the reduction of PDK4 level in the PSEsusceptible turkeys, PDH continues converting pyruvate into acetyl CoA without the inhibiting signal from PDK4, resulting in accelerated rate of glucose oxidation. Oxygen is rapidly consumed. Metabolism switches from aerobic into anaerobic pathways faster in the susceptible birds, resulting in greater amounts of lactic acid accumulated in the muscle cell that causes rapid postmortem pH decline in the defective meat. In addition, as increased ATP may be generated in the + susceptible turkeys, H production during hydrolysis of ATP may accelerate pH drop and augment the acidosis in the susceptible birds. Based on the hypothesis of altered glucose oxidation in PSE turkey, ATP levels in PSE turkey may be greater than that of normal samples as ATP is continuously generated via glucose oxidation early postmortem, and rapidly utilized for cellular activities afterward. Rapid ATP degradation was previously observed in skeletal muscle of pigs carrying heterozygous halothane gene, the gene associated with lean muscle, but pigs carrying this gene potentially yield PSE meat (Lahucky et al., 2002; Moesgaard et al., 1995; Shen et al., 2007). However, in those studies, the ATP content at the beginning of the time-course experiments was not significantly different between halothane-gene carriers and ! 88! non-carriers. In addition, Miri et al. (1992) and Batlle et al. (2000) reported lower postmortem ATP content in PSE pork. However, Miri et al. and Batlle et al. quantified ATP at 30 min postmortem and 2h postmortem, which might be too metabolically advanced and thus they were not able to capture the rapid ATP metabolism at the onset of postmortem conversion of muscle to meat. Comparison of ATP content and its degradation rate between normal and PSE turkey at the onset of postmortem will provide insight regarding the impact of down-regulated PDK4 on glucose oxidative metabolism. Knowledge of the mechanisms regulating PDK4 expression is beneficial for developing pathway intervention approaches or nutritional modification to minimize the incidence of PSE turkey. Expression of PDK4 is suppressed by insulin and induced by glucocorticoids (Connaughton et al, 2010) as well as highfat feed (Holness et al., 2000). Increase in PDK4 protein level was also found during starvation (Wu et al., 2000). Above all, better understanding of PDK4 roles on development of PSE meat defect is needed so that the most effective approaches can be carefully developed and implemented without creating other problems for the industry in the future. 4.5 Conclusions A significant decrease in PDK4 mRNA and protein was identified in both modern and random-bred turkey lines. This observation supports the hypothesis of an abnormality in the glucose oxidative pathway, which may result in unusually high rate of postmortem metabolism in PSE turkey. Comparison of PDK4 and ! 89! PDH activity, as well as oxygen consumption rate, between normal and PSE turkey will provide better confirmation of the hypothesis of glucose oxidative dysregulation and better comprehension on PDK4 biological roles on development of this turkey meat defect. ! 90! CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTIONS Differential gene expression between turkey skeletal muscle samples of normal and PSE turkey meat collected from the randombred control line (RBC2) has been revealed by transcriptome analyses. In the global gene expression analysis using the turkey skeletal muscle long oligonucleotide (TSKMLO) microarray, 49 transcripts were reported with significant difference in expression at FDR < 0.1. Deep transcriptome sequencing analysis (RNA-Seq) identified 494 differentially expressed transcripts between normal and PSE samples at FDR < 0.05. Pathway analysis strongly supports the association of multiple cellular mechanisms in development of the PSE turkey meat defect. A clear association between an abnormality of Ca 2+ homeostasis and development of this turkey meat defect has been confirmed. In addition, upregulation of the gene encoding the delta subunit of the skeletal muscle nicotinic acetylcholine receptor was identified in PSE turkey, suggesting dysregulation of 2+ [Ca ] in PSE turkey at both intracellular and extracellular levels. Actin cytoskeletons of the PSE-susceptible birds are affected as multiple signaling pathways regulating formation and reorganization of actin cytoskeleton are altered. Irregular actin cytoskeleton-related pathways might affect various downstream cellular activities as well as structural stability of muscle cells. Muscle cells of the susceptible birds may be intolerant of a drastic change of muscle shape and volume during extensive muscle contraction; hence, muscle damage, ! 91! which detrimentally affects quality of the meat. Additionally, down-regulation of genes encoding fast-twitch muscle isoforms and down-regulation of slow-twitch isoforms may imply fast-to-slow switch muscle isoform conversion. The change in ratio of muscle isoforms and the other skeletal muscle regulatory proteins may interfere with biochemical activities inside the cells. Changes in biochemical activities either from irregular actin cytoskeleton-mediated pathways or from transformation of muscle isoforms may manifest the process of conversion of muscle to meat, leading to the inferior characteristics of PSE turkey. Based on the reduction of both mRNA and protein abundance of pyruvate dehydrogenase kinase isozyme 4 (PDK4) in PSE turkey from the randombred line, we propose a new postmortem paradigm regarding development of PSE turkey. The rapid postmortem pH drop in the susceptible birds may not be only from high rate of glycolysis, but also from an altered regulation of glucose oxidative metabolism. At the early postmortem stage when oxygen is still present, muscle cells generate energy for cellular activities via glucose oxidative metabolism. Subsequently, when oxygen is completely consumed, the pathways switch to anaerobic pathways in which lactic acid is produced, causing pH decline in the muscle. However, in the PSE-susceptible turkeys, a decrease in PDK4 level accelerates conversion of pyruvate into acetyl CoA. Oxygen is consumed more rapidly, leading to a faster rate of postmortem metabolic switch from aerobic to anaerobic pathways. A large amount of lactic acid accumulation combined with protons generated from ATP hydrolysis results in rapid pH decline. The biological relevance of decreased PDK4 expression and development of PSE turkey is ! 92! strongly supported by down-regulation of PDK4 transcript and protein abundance in PSE meat of the modern commercial line in which more frequent prevalence and greater magnitude of the PSE problem has been observed. Further investigation is recommended to test the proposed hypothesis. The study may include comparison of PDK4 activity between normal and PSE turkey. This can be indirectly indicated by determining protein abundance of pyruvate dehydrogenase (PDH) and its phosphorylated forms using immunoassays. Lower phosphorylated PDH abundance is anticipated in PSE samples. In addition, a study designed to follow interval postmortem ATP change in turkey meat samples will test the hypothesis of an abnormality of glucose oxidative metabolism in PSE turkey. A set of meat samples from each biological replicate can be collected at different postmortem time-points and immediately snap frozen, hence each sample provides a snapshot of metabolic information at a definite postmortem period. It can be presumed that, once all of the time-point ATP metabolic data is combined, the dataset will depict overall postmortem ATP metabolism in turkey meat. The quantity of ATP can be determined in intact meat samples using 31 phosphorus-31 nuclear magnetic resonance spectroscopy ( P-NMR); the more sensitive and more rapid non-destructive technique compared with other biochemical analytical methods. The speculated outcome is that ATP level in PSE turkey is greater than that of normal samples at early postmortem stage, as it is continuously generated via glucose oxidation, and rapidly declines afterward. However, there are some difficulties to be recognized, including how quickly meat ! 93! samples must be collected after slaughtering, and what optimized conditions to operate 31 P-NMR. It must be noted that PDK4 received priority for an investigation at the protein level in this dissertation. This is due to its substantial down-regulation reported by both transcriptome analyses as well as its biological relevance to development of PSE meat. Other candidate genes can be selected from the other associated differentially expressed genes using similar criteria to confirm the different expression at protein level and explore its biological manifestation in development of the turkey meat defect. Collectively, we are one step closer to obtaining a complete understanding of molecular mechanisms underlying development of PSE turkey. Pathway intervention approaches can be developed and implemented to minimize occurrence of PSE turkey. For example, based on the findings in this dissertation, an induced expression of PDK4 is in response of low blood glucose level and glucocorticoids secretion. The use of nutrient modification by reducing percentage of carbohydrate and increasing protein in feed may induce expression of PDK4 and compensate for the impairment of glucose oxidation. Further tests in the turkey cell lines as well as animal subjects are encouraged before an actual application with the population to ensure that the approaches are effective and less likely to generate other problems for the industry. It is also recommended to determine transcriptome profiles of normal and PSE turkey subjected to thermal stress. An increased occurrence of “light meat” ! 94! with inferior quality characteristics has been reported when turkeys are exposed to heat stress. As the global average temperature has been rising, weather anomalies, including heat waves and harsh cold temperatures will occur more frequently. Exposure to weather extremes can induce thermal stress in the animals. The fast-growing turkeys display less environmental heat tolerance; thus, the modern commercial birds may struggle with the thermal challenge due to the climate changes and produce more PSE meat. Conducting a series of studies similar to the current study will provide mechanistic comprehension of the PSE problem in the heat stress exposed turkeys. The biochemical information will be beneficial for the development of PSE-intervention approaches that work effectively in both heat-stress and non-heat stress populations. ! 95! APPENDICES ! 96! APPENDIX A CLASSIFICATION OF PALE, SOFT AND EXUDATIVE TURKEY MEAT The problem of PSE meat in the poultry industry has been one of the greatest challenges to both producers and processors. Meat processors contend with undesirable characteristics of processed products made from the PSE meat due to denaturation of myofibrillar proteins. Separation of PSE meat from normal meat before the defective meat enters the process is the most effective strategy to minimize defective products from PSE meat in commercial plants (reviewed by Barbut et al., 2008). However, establishment of sorting criteria precisely representing protein denaturation is still problematic for both industry and scientific investigators. Previous studies proposed to utilize meat color as a sorting indicator for PSE poultry meat (Barbut 1998) as it can be measured easily and rapidly with a non-destructive method in the processing line. In the poultry industry, color of the meat was originally determined by a visual scoring system, and this method has continued in some plants (Barbut 2002). However, this method is effective only when identifying PSE meat with an extreme discoloration. Poultry PSE is difficult to differentiate from normal meat as the pale appearance of poultry PSE meat is not as visually obvious as in porcine PSE meat due to the differences in muscle structure and composition between species (Aberle et al., 2001). Measurement of meat color using colorimetric devices has been the most popular method used by scientific investigators and also implemented in some ! 97 industrial plants. Colorimetric equipment provides more accurate readings and overcomes the inherent human-error of visual scoring (Barbut, 2002). In the color system, lightness correlates with an increasing L* value (range from 0 to 100); therefore, the greater the L* value, the paler the meat is. Development of L* cutoff appears to be a promising indicator in classifying poultry PSE meat for both industry and scientific study. However, there is inconsistency among previous reports as to what L* cutoff should be exploited (Table A.1). Previous investigators also observed that lightness of meat as reported by L* considerably varies in accordance with many factors, including transportation (Bianchi et al., 2005), and color measuring practices (Barbut, 1997a; Bianchi et al., 2002; Garcia et al., 2010; Petracci and Fletcher, 2002; Wilkins et al., 2000), as well as seasons (Barbut, 1997b; Bianchi et al., 2007; McCurdy et al., 1996; Petracci et al., 2004). Color variation is also found among birds from different genetic lines (Janisch et al., 2011), genders (Chiang et al., 2008) and flocks (Barbut, 1998). Based on previous research, it is difficult to define a stringent L* cutoff point for classifying poultry PSE meat. ! 98 Table A.1 Cutoff values used in previous literature for classifying poultry PSE meat Barbut 1996 Turkey Cutoff value PSE Normal (N/A) L* > 50 McKee and Sams 1997 Turkey L* > 53 (N/A) Pietrzak et al. 1997 Turkey pH20 min ' 5.8 Barbut 1998 Turkey Chicken L* ( 53 pH20 min > 5.8 (N/A) (N/A) Investigators Published Sample year L* ( 50 Owens et al. 2000 Turkey L* ( 53 L* < 53 Petracci et al. 2004 Chicken L* > 56 50 ' L* ' 56 Barbut et al. 2005 Chicken L* > 53 pH < 5.7 46 < L* ' 53 5.7 < pH < 6.1 Bianchi et al. 2005 Chicken L* > 53 46 < L* ' 53 Zhang and Barbut 2005 Chicken L* > 53 pH < 5.7 46 < L* < 53 5.9 < pH < 6.1 Fraqueza et al. 2006 Turkey L* ( 50 40 < L* < 50 pH24 h < 5.8 Marchi et al. Chicken L* ( 53 pH < 5.9 Oda et al. 2009 Chicken L* > 53 44 ' L* ' 53 pH ( 5.9 (N/A) Garcia et al. 2010 Chicken L* ( 49 (N/A) Gorsuch and Alvarado Wilhelm et al. 2010 Chicken L* > 53 L* < 53 2010 Chicken pH1.5 h < 6.0 pH1.5 h > 6.0 Ziober et al. 2010 Chicken L* > 53 44 < L*< 53 Droval et al. 2012 Chicken pH1.5 h < 6.0 Samuel et al. 2012 Chicken 2012 Turkey L* > 60 Low %marinade uptake High %cook loss pH1.5 h > 6.0 (N/A) Sporer et al. High %marinade uptake Low %cook loss Zhu et al. ! 2009 2012 Chicken L* > 53 48 < L*' 53 99 Our group also experienced variation of meat lightness among turkey samples from different genetic lines (Figure A.1). The turkey breasts of randombred control (RBC2) line birds were darker than those of the modern commercial (COMM) line. If the samples were classified as PSE when L* > 53, the cutoff generally used in most previous research, none of the samples from the COMM line would be considered “normal” samples. In addition, lightness of the meat samples from both lines showed no significant correlation with marinade uptake (P ( 0.05, Figure A.2). For cook loss, weak positive correlation between lightness and cook loss was found in RBC2 turkey meat (P < 0.05), while no significant correlation was found in COMM samples (P ( 0.05). High marinade uptake and low cook loss represent superior functionality, particularly water holding capacity, of myofibrillar proteins. The results implied that L* value may not be an accurate predictor of protein functionality which agrees with the observation of Samuel et al. (2012) reporting a weak correlation between lighter color of poultry meat and water holding capacity. ! 100 (a) RBC2 (n = 43) (b) COMM (n = 39) Number of meat samples Figure A.1 Color distribution of turkey breast meat. Meat samples were collected from (a) the randombred control line (RBC2, n = 43) or (b) the modern commercial line (COMM, n = 39). ! 101 (a) Uptake (%) = -0.64 (L*) + 65.5 (b) Cook loss (%) = 0.28 (L*) + 14.7 2 2 R = 0.01, P = 0.52 R = 0.12, P = 0.02 (c) Uptake (%) = -0.72(L*) + 56.9 (d) Cook loss (%) = 0.01(L*) + 29.1 2 2 R = 0.04, P = 0.22 R = 4.9E-05, P = 0.96 Figure A.2 Correlations between lightness (L*) and meat quality indices (percent marinade uptake or percent cook loss) of turkey breasts. Samples were collected from randombred control (RBC2) turkeys (n = 43) and modern commercial (COMM) turkeys (n = 39). (a, b) Correlations between L* and meat quality of RBC2 turkeys. (c, d) Correlations between L* and meat quality of COMM turkeys. ! 102 Based on weak correlations of L* value and indices for protein functionality, our group decided to use extremes of percentage of marinade uptake and cook loss as primary and secondary indices to classify turkey breast meat as normal or PSE samples within each treatment group (Sporer et al., 2012). The poor water holding capacity of myofibrillar proteins of raw and cooked meat can be indirectly indicated by low marinade uptake and high cook loss. Therefore, the samples with this characteristic would be classified as PSE samples. Conversely, high marinade uptake and low cook loss indicate better protein functionality of myofibrillar proteins; thus, they would be classified as normal samples. We argue that the classification of samples by these criteria provide better representation for this study, which focuses on comparison of gene expression patterns and molecular functions between sample groups with PSE or normal meat quality. ! 103 APPENDIX B DIFFERENTIALLY EXPRESSED TRNASCRIPTS IDENTIFIED BY RNA-SEQ ANALYSIS (FDR<0.05) Table B.1 Differentially expressed transcript identified in RNA-Seq (false discovery rate; FDR<0.05) Reference gene name arylacetamide deacetylase (esterase) 4-aminobutyrate aminotransferase ATP-binding cassette, sub-family A (ABC1), member 5 abhydrolase domain containing 5 actin binding LIM protein family, member 3 acetyl-CoA acyltransferase 2 FC -2.1 -1.9 -2.2 1 FDR 9.41E-06 2.38E-04 1.41E-04 -1.9 1.8 1.6 6.73E-04 1.73E-02 1.87E-02 acetyl-coenzyme A carboxylase beta acyl-CoA binding domain containing 7 angiotensin I converting enzyme (peptidyldipeptidase A) 1 -1.5 -2.3 2.0 4.76E-02 3.40E-03 1.40E-02 ACOT11 ACSBG2 ACTA2 ACTC1 ACTG2 ADAMTS 12 acyl-CoA thioesterase 11 acyl-CoA synthetase bubblegum family member 2 actin, alpha 2, smooth muscle, aorta actin, alpha, cardiac muscle 1 actin, gamma 2, smooth muscle, enteric ADAM metallopeptidase with thrombospondin type 1 motif, 12 -4.0 1.7 1.9 7.2 3.0 2.3 1.69E-17 1.53E-02 6.95E-04 1.29E-33 1.12E-06 1.04E-04 ADAMTS 15 ADAM metallopeptidase with thrombospondin type 1 motif, 15 2.5 5.87E-07 ADAMTS 2 ADAM metallopeptidase with thrombospondin type 1 motif, 2 1.9 8.31E-04 ADAMTS 9 ADAM metallopeptidase with thrombospondin type 1 motif, 9 1.8 3.20E-03 ADIPOQ ADIPOR2 ADPRHL 1 AEBP1 adiponectin, C1Q and collagen domain containing adiponectin receptor 2 ADP-ribosylhydrolase like 1 1.9 -2.0 -1.6 2.41E-04 1.31E-04 3.02E-02 AE binding protein 1 1.8 6.61E-03 Gene ID AADAC ABAT ABCA5 ABHD5 ABLIM3 ACAA2 ACACB ACBD7 ACE ! 104 Table B.1 (cont’d) Gene ID AGBL1 AGPAT9 AGRN AMD1 ANGPT4 ANGPTL 1 ANKRD2 4 ANO6 ANXA1 ANXA2 ANXA7 APP ARL4C ARRDC2 ASB2 ASB4 ASNS ASPA ATF3 ATOH8 ATP1B4 ATP5I B1N1B8_ MELGA B1N1B8_ MELGA B3VL40_ MELGA BBOX1 BCAT1 ! Reference gene name ATP/GTP binding protein-like 1 1-acylglycerol-3-phosphate O-acyltransferase 9 agrin adenosylmethionine decarboxylase 1 angiopoietin 4 angiopoietin-like 1 FC 2.5 -1.7 2.1 1.9 2.0 2.4 FDR 3.07E-09 6.63E-03 5.19E-03 3.50E-04 3.93E-03 4.31E-02 ankyrin repeat domain 24 10.4 -1.6 2.1 1.5 1.8 1.9 3.3 -6.5 -2.2 -1.8 2.1 -2.7 1.6 -2.3 -5.2 1.5 1.40E-15 -3.7 8.38E-09 Uncharacterized protein -2.8 1.51E-07 delta-like 1 homolog (Drosophila) -2.0 3.42E-02 butyrobetaine (gamma), 2-oxoglutarate dioxygenase (gamma-butyrobetaine hydroxylase) 1 branched chain amino-acid transaminase 1, cytosolic -1.7 1.32E-02 2.3 5.18E-06 anoctamin 6 annexin A1 annexin A2 annexin A7 amyloid beta (A4) precursor protein ADP-ribosylation factor-like 4C arrestin domain containing 2 ankyrin repeat and SOCS box containing 2 ankyrin repeat and SOCS box containing 4 asparagine synthetase (glutamine-hydrolyzing) aspartoacylase activating transcription factor 3 atonal homolog 8 (Drosophila) ATPase, Na+/K+ transporting, beta 4 polypeptide ATP synthase, H+ transporting, mitochondrial Fo complex, subunit E Uncharacterized protein 105 2.73E-02 9.04E-04 2.83E-02 1.05E-02 1.09E-03 1.45E-02 6.71E-34 9.40E-06 1.19E-03 4.88E-04 3.14E-04 1.57E-02 4.89E-05 5.48E-25 3.62E-02 Table B.1 (cont’d) Gene ID BCHE BDH1 BEST3 BMF BOC C10orf26 C13orf39 C1orf96 C1QTNF 1 C1QTNF 2 C1S C2CD3 C2orf40 CA4 CACNG4 Reference gene name butyrylcholinesterase 3-hydroxybutyrate dehydrogenase, type 1 bestrophin 3 Bcl2 modifying factor Boc homolog (mouse) WW domain binding protein 1-like methyltransferase like 21C chromosome 1 open reading frame 96 C1q and tumor necrosis factor related protein 1 FC 3.3 -1.8 -3.2 -2.7 1.7 -2.3 2.2 -2.3 3.6 FDR 6.43E-03 5.19E-03 2.86E-12 1.00E-08 3.45E-02 5.07E-07 8.81E-07 4.09E-02 2.34E-03 C1q and tumor necrosis factor related protein 2 2.2 3.42E-04 complement component 1, s subcomponent C2 calcium-dependent domain containing 3 chromosome 2 open reading frame 40 carbonic anhydrase IV calcium channel, voltage-dependent, gamma subunit 4 1.8 -2.8 2.6 4.0 2.3 4.18E-02 1.11E-03 1.55E-03 7.72E-09 1.50E-04 CARNS1 CAT CCDC13 5 CCDC14 1 CCDC80 CCPG1 CCRN4L carnosine synthase 1 catalase coiled-coil domain containing 135 -1.5 -2.5 -1.6 4.09E-02 2.79E-08 4.19E-02 coiled-coil domain containing 141 -1.9 1.77E-04 coiled-coil domain containing 80 cell cycle progression 1 CCR4 carbon catabolite repression 4-like (S. cerevisiae) 1.8 -1.7 -2.3 1.99E-03 2.52E-03 3.19E-04 CD34 molecule CD59 molecule, complement regulatory protein cadherin 11, type 2, OB-cadherin (osteoblast) centromere protein V 1.6 1.7 1.7 1.7 4.04E-02 2.66E-02 4.19E-02 1.35E-02 1.7 2.8 4.72E-02 1.49E-08 -1.5 4.64E-02 CD34 CD59 CDH11 CENPV CERCAM cerebral endothelial cell adhesion molecule CHAD chondroadherin CHMP6 ! charged multivesicular body protein 6 106 Table B.1 (cont’d) Gene ID CHODL CHRND CIDEC CILP Reference gene name chondrolectin cholinergic receptor, nicotinic, delta cell death-inducing DFFA-like effector c cartilage intermediate layer protein, nucleotide pyrophosphohydrolase CLEC3B CLTA CLU CNR1 CNTN4 COL16A1 COL1A2 COL20A1 COL21A1 COL28A1 COL3A1 COL4A6 COL5A1 COL5A2 COL6A1 COL6A2 COL6A3 COQ10B CPE CPEB4 C-type lectin domain family 3, member B clathrin, light chain A clusterin cannabinoid receptor 1 (brain) contactin 4 collagen, type XVI, alpha 1 collagen, type I, alpha 2 collagen, type XX, alpha 1 collagen, type XXI, alpha 1 collagen, type XXVIII, alpha 1 collagen, type III, alpha 1 collagen, type IV, alpha 6 collagen, type V, alpha 1 collagen, type V, alpha 2 collagen, type VI, alpha 1 collagen, type VI, alpha 2 collagen, type VI, alpha 3 coenzyme Q10 homolog B (S. cerevisiae) carboxypeptidase E cytoplasmic polyadenylation element binding protein 4 CREB3L1 cAMP responsive element binding protein 3-like 1 CREB5 cAMP responsive element binding protein 5 CRYAB crystallin, alpha B CUBN cubilin (intrinsic factor-cobalamin receptor) CXCL12 chemokine (C-X-C motif) ligand 12 CXCL14 chemokine (C-X-C motif) ligand 14 ! 107 FC 2.2 3.4 -2.5 -2.2 FDR 2.54E-02 6.99E-13 4.09E-02 1.73E-05 2.0 1.8 1.6 2.4 2.2 2.2 3.9 3.9 2.5 1.7 3.3 2.0 2.7 2.0 2.8 2.6 1.7 -1.7 2.7 -1.6 6.72E-05 3.58E-02 1.82E-02 5.98E-05 1.25E-02 1.31E-03 7.89E-17 2.49E-05 5.16E-03 1.06E-02 2.35E-08 3.88E-02 1.26E-09 1.44E-04 1.92E-08 7.66E-09 8.96E-03 1.00E-02 3.54E-03 3.09E-02 1.7 2.0 -2.0 1.8 2.0 2.0 2.18E-02 8.71E-06 5.35E-05 2.83E-02 1.90E-04 3.19E-02 Table B.1 (cont’d) Gene ID DACT1 Reference gene name dapper, antagonist of beta-catenin, homolog 1 (Xenopus laevis) DBX1 developing brain homeobox 1 DCBLD1 discoidin, CUB and LCCL domain containing 1 DCN decorin DCTN3 dynactin 3 (p22) DDB2 damage-specific DNA binding protein 2, 48kDa DDR2 discoidin domain receptor tyrosine kinase 2 DEGS1 degenerative spermatocyte homolog 1, lipid desaturase (Drosophila) DENND2 DENN/MADD domain containing 2C C FC 1.8 FDR 2.83E-02 2.3 2.0 1.9 1.6 1.7 2.5 1.8 3.03E-03 1.37E-03 3.61E-04 1.60E-02 2.52E-02 7.24E-04 8.52E-04 -1.7 7.65E-04 DHX58 DNAJA4 DNASE1 L3 DUSP26 DUSP8 ECM2 -1.9 -1.7 -2.2 2.54E-02 1.61E-02 3.31E-04 dual specificity phosphatase 26 (putative) dual specificity phosphatase 8 extracellular matrix protein 2, female organ and adipocyte specific -2.0 -1.8 1.8 3.25E-03 5.82E-04 7.90E-03 EDA2R EGF EGR1 EHHAD H ectodysplasin A2 receptor epidermal growth factor early growth response 1 enoyl-CoA, hydratase/3-hydroxyacyl CoA dehydrogenase 2.1 1.6 1.8 -2.2 1.25E-05 1.01E-02 7.90E-03 3.63E-02 EIF2AK3 eukaryotic translation initiation factor 2-alpha kinase 3 1.8 1.98E-02 ELN EMILIN2 ENAH EPHB1 EPN3 ESR1 ETS2 ! DEXH (Asp-Glu-X-His) box polypeptide 58 DnaJ (Hsp40) homolog, subfamily A, member 4 deoxyribonuclease I-like 3 14.3 1.7 1.6 1.7 -2.5 -2.1 1.6 1.20E-55 7.11E-03 3.77E-03 6.68E-03 5.64E-06 8.06E-04 3.37E-02 elastin elastin microfibril interfacer 2 enabled homolog (Drosophila) EPH receptor B1 epsin 3 estrogen receptor 1 v-ets erythroblastosis virus E26 oncogene homolog 2 (avian) 108 Table B.1 (cont’d) Gene ID FAM117 A FAM13A FAM173 B FAM46A FAM55C FAM70A FBLN1 FBLN2 FBP2 FBXO32 Reference gene name family with sequence similarity 117, member A FC -1.7 FDR 3.94E-03 family with sequence similarity 13, member A family with sequence similarity 173, member B -3.0 -2.1 1.41E-12 4.98E-03 family with sequence similarity 46, member A family with sequence similarity 55, member C family with sequence similarity 70, member A fibulin 1 fibulin 2 fructose-1,6-bisphosphatase 2 F-box protein 32 1.8 -2.3 -2.1 1.9 1.7 1.6 -6.5 9.38E-03 2.66E-06 5.28E-05 3.25E-03 8.60E-03 1.55E-02 5.10E-32 FGFR1 FHL1 FHL2 FHOD3 FKBP10 FKBP5 FKTN FMOD FN1 FNDC1 FNDC5 FSCN1 fibroblase growth factor receptor 1 four and a half LIM domains 1 four and a half LIM domains 2 orming homology 2 domain containing 3 FK506 binding protein 10, 65 kDa FK506 binding protein 5 fukutin fibromodulin fibronectin 1 fibronectin type III domain containing 1 fibronectin type III domain containing 5 fascin homolog 1, actin-bundling protein (Strongylocentrotus purpuratus) 1.6 1.9 2.7 1.8 2.5 -2.5 2.9 5.4 1.9 2.8 2.2 2.3 4.70E-02 5.92E-05 1.92E-05 8.41E-04 3.64E-06 2.59E-08 3.86E-12 2.03E-15 1.07E-03 9.72E-09 3.39E-03 9.20E-04 FSTL1 FYN follistatin-like 1 FYN oncogene related to SRC, FGR, YES 2.1 2.0 3.69E-05 5.96E-04 1.7 1.9 4.90E-02 3.50E-02 -1.5 4.10E-02 1.6 3.5 2.58E-02 6.44E-04 G0S2 G0/G1 Switch protein 2 GABRA4 gamma-aminobutyric acid (GABA) A receptor, alpha 4 GADD45 growth arrest and DNA-damage-inducible, alpha A GADL1 glutamate decarboxylase-like 1 GALNT5 UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 5 (GalNAc-T5) ! 109 Table B.1 (cont’d) Gene ID Reference gene name GALNTL1 UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase-like 1 GATM glycine amidinotransferase (L-arginine:glycine amidinotransferase) GDAP1 ganglioside-induced differentiation-associated protein 1 GDF8_M Growth/differentiation factor 8 ELGA GGT5 gamma-glutamyltransferase 5 GJA1 gap junction protein, alpha 1, 43kDa GLRX glutaredoxin (thioltransferase) GLUL glutamate-ammonia ligase GPRC5C G protein-coupled receptor, family C, group 5, member C GRB10 GRK7 GTSF1 GYLTL1B HAPLN3 HBA_ME LGA HBAD_M ELGA HDAC9 HECTD2 HIC1 HTRA3 IGFBP3 IGFBP4 IGFBP7 IGFN1 IGJ IGSF3 IL13RA2 ! FC 2.8 FDR 1.33E-04 2.6 6.81E-09 -2.0 1.75E-02 2.1 5.76E-05 2.1 2.0 1.7 -2.5 -2.5 1.41E-02 2.20E-04 3.76E-03 1.36E-09 6.10E-04 growth factor receptor-bound protein 10 G protein-coupled receptor kinase 7 gametocyte specific factor 1 glycosyltransferase-like 1B hyaluronan and proteoglycan link protein 3 Uncharacterized protein -2.2 4.1 -3.3 1.6 2.1 1.8 3.46E-05 8.02E-06 2.95E-06 3.16E-02 7.94E-03 2.21E-03 Uncharacterized protein 2.8 1.64E-06 histone deacetylase 9 HECT domain containing 2 hypermethylated in cancer 1 hypermethylated in cancer 1 insulin-like growth factor binding protein 3 insulin-like growth factor binding protein 4 insulin-like growth factor binding protein 7 immunoglobulin-like and fibronectin type III domain containing 1 immunoglobulin J polypeptide, linker protein for immunoglobulin alpha and mu polypeptides immunoglobulin superfamily, member 3 interleukin 13 receptor, alpha 2 1.7 -2.3 1.6 1.9 1.6 1.8 1.7 -1.9 8.19E-03 4.01E-04 2.70E-02 4.16E-03 4.11E-02 7.85E-03 9.04E-03 3.17E-02 4.0 6.50E-08 1.6 5.1 2.06E-02 4.16E-03 110 Table B.1 (cont’d) Gene ID INSIG1 IRF8 ISM1 ISM2 ITGA11 ITGA6 ITM2C JAM3 KAZALD1 KBTBD5 KCNC1 KCNIP2 KCTD20 KIAA0408 KIRREL KLF15 KLHL23 KLHL24 KLHL38 KY LAMP2 LEPREL4 LGALS1 LGI2 LIMA1 LINGO1 LIPA LMO4 LRFN2 LRRC17 LRRC8A ! Reference gene name insulin induced gene 1 interferon regulatory factor 8 isthmin 1 homolog (zebrafish) isthmin 2 homolog (zebrafish) FC 1.8 -1.6 4.8 1.9 FDR 2.29E-03 3.85E-02 1.84E-04 7.01E-04 integrin alpha 11 integrin, alpha 6 integral membrane protein 2C junctional adhesion molecule 3 Kazal-type serine peptidase inhibitor domain 1 kelch repeat and BTB (POZ) domain containing 5 potassium voltage-gated channel, Shaw-related subfamily, member 1 1.8 2.1 -2.1 2.5 3.2 1.8 2.4 4.70E-02 7.53E-05 3.06E-05 3.35E-06 4.69E-06 2.17E-03 3.28E-04 Kv channel interacting protein 2 potassium channel tetramerisation domain containing 20 2.8 -1.7 8.38E-06 7.19E-03 KIAA0408 kin of IRRE like (Drosophila) Kruppel-like factor 15 Uncharacterized protein Uncharacterized protein Uncharacterized protein kyphoscoliosis peptidase lysosomal-associated membrane protein 2 leprecan-like 4 lectin, galactoside-binding, soluble, 1 leucine-rich repeat LGI family, member 2 LIM domain and actin binding 1 leucine rich repeat and Ig domain containing 1 lipase A, lysosomal acid, cholesterol esterase LIM domain only 4 leucine rich repeat and fibronectin type III domain containing 2 leucine rich repeat containing 17 leucine rich repeat containing 8 family, member A -2.6 1.6 -2.2 -1.9 -2.6 -21.5 1.6 -1.7 2.3 2.2 1.7 1.7 -1.7 1.7 1.6 2.8 1.50E-07 1.92E-02 3.99E-06 1.25E-02 4.22E-09 2.54E-59 3.44E-02 2.60E-03 1.27E-02 2.24E-04 2.73E-02 4.32E-02 1.25E-02 4.39E-02 3.77E-02 1.11E-03 2.0 1.7 3.94E-03 7.53E-03 111 Table B.1 (cont’d) Gene ID LTBP1 Reference gene name latent transforming growth factor beta binding protein 1 FC 1.8 FDR 4.47E-03 1.7 4.59E-03 2.3 3.6 -1.8 1.03E-06 3.64E-03 3.75E-02 MAM domain containing 2 microtubule-associated protein 1B mitogen-activated protein kinase kinase 6 mitogen-activated protein kinase kinase kinase 15 mitogen-activated protein kinase kinase kinase 4 microtubule-associated protein, RP/EB family, member 2 membrane bound O-acyltransferase domain containing 2 2.5 -2.6 -1.7 -1.8 1.7 -2.0 6.33E-06 1.26E-05 7.98E-03 5.07E-03 2.80E-04 1.93E-04 2.2 4.99E-03 MCL1 MFAP5 MLF1 MMP7 MRC2 MRPL50 MTAP MTHFD2 myeloid cell leukemia sequence 1 (BCL2-related) microfibrillar associated protein 5 myeloid leukemia factor 1 matrix metallopeptidase 7 (matrilysin, uterine) mannose receptor, C type 2 mitochondrial ribosomal protein L50 methylthioadenosine phosphorylase methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 2, methenyltetrahydrofolate cyclohydrolase -1.8 1.7 1.5 Inf 1.7 1.6 1.6 2.2 6.06E-03 1.58E-02 3.59E-02 1.79E-03 3.30E-03 3.52E-02 3.80E-02 3.92E-06 MTMR3 MXI1 MXRA8 MYBPC1 MYH10 MYH11 MYH13 MYL2 myotubularin related protein 3 MAX interactor 1 matrix-remodelling associated 8 myosin binding protein C, slow type myosin, heavy chain 10, non-muscle myosin, heavy chain 11, smooth muscle myosin, heavy chain 13, skeletal muscle myosin, light chain 2, regulatory, cardiac, slow -12.4 -2.1 1.9 1.9 1.5 2.3 -4.9 1.6 4.58E-29 2.38E-05 2.19E-03 4.76E-04 4.19E-02 3.27E-07 1.01E-17 1.26E-02 LTBP2 latent transforming growth factor beta binding protein 2 LUM lumican MAB21L1 mab-21-like 1 (C. elegans) MAGI1 membrane associated guanylate kinase, WW and PDZ domain containing 1 MAMDC2 MAP1B MAP2K6 MAP3K15 MAP3K4 MAPRE2 MBOAT2 ! 112 Table B.1 (cont’d) Gene ID MYL9 MYLK_M ELGA MYLK4 MYO9A NDRG4 NETO2 NFKBIA NOTUM NOV NRBP1 NRP1 NRXN1 NTN4 NTSR1 NUAK1 OAF OGN OIT3 OLFML2 B OSR1 P84479_ MELGA PAOX PAQR7 PARP3 PCMTD1 PDCD4 PDE10A PDGFRB ! Reference gene name myosin, light chain 9, regulatory Uncharacterized protein FC 2.0 1.8 FDR 1.11E-03 6.61E-03 myosin light chain kinase family, member 4 myosin IXA NDRG family member 4 neuropilin (NRP) and tolloid (TLL)-like 2 nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha notum pectinacetylesterase homolog nephroblastoma overexpressed gene nuclear receptor binding protein 1 neuropilin 1 neurexin 1 netrin 4 neurotensin receptor 1 (high affinity) NUAK family, SNF1-like kinase, 1 OAF homolog (Drosophila) osteoglycin oncoprotein induced transcript 3 olfactomedin-like 2B -4.4 1.6 1.8 3.2 -1.6 1.99E-20 4.49E-02 2.18E-02 1.48E-09 1.83E-02 1.7 37.7 1.6 1.7 1.6 -1.9 1.9 -1.7 2.2 2.5 2.3 1.7 2.69E-02 1.65E-64 3.63E-02 1.41E-02 3.19E-02 3.12E-02 1.04E-03 2.63E-02 1.46E-04 4.65E-07 6.55E-03 7.05E-03 odd-skipped related 1 (Drosophila) Uncharacterized protein 2.5 1.8 1.23E-02 3.88E-03 polyamine oxidase (exo-N4-amino) progestin and adipoQ receptor family member VII poly (ADP-ribose) polymerase family, member 3 protein-L-isoaspartate (D-aspartate) Omethyltransferase domain containing 1 -2.1 -3.3 -2.0 -1.8 9.06E-03 3.62E-06 4.05E-04 1.21E-03 programmed cell death 4 (neoplastic transformation inhibitor) -2.0 2.33E-04 phosphodiesterase 10A platelet-derived growth factor receptor, beta polypeptide -5.8 1.9 1.39E-19 6.73E-04 113 Table B.1 (cont’d) Gene ID PDGFRL PDK4 PFKFB3 FC 2.4 -14.1 -4.2 FDR 1.23E-04 9.84E-56 1.95E-16 1.8 3.92E-03 2.2 -1.7 -4.2 7.80E-06 3.30E-03 2.11E-02 -1.7 5.38E-03 phospholipase C, delta 4 -1.6 3.26E-02 PLEKHN 1 PLIN1 PLK2 pleckstrin homology domain containing, family N member 1 perilipin 1 polo-like kinase 2 1.7 1.82E-02 -1.7 2.1 1.53E-02 3.09E-03 PLXND1 PM20D2 PMM1 POSTN PPAP2B PPARD PPIC PPP2R2 C plexin-D1 peptidase M20 domain containing 2 phosphomannomutase 1 periostin, osteoblast specific factor phosphatidic acid phosphatase type 2B peroxisome proliferator-activated receptor delta peptidylprolyl isomerase C (cyclophilin C) protein phosphatase 2, regulatory subunit B, gamma 1.5 -1.6 -1.9 4.1 1.6 -2.0 3.3 -2.3 4.84E-02 2.38E-02 1.59E-03 9.76E-18 2.26E-02 4.07E-03 1.52E-08 5.38E-03 PPTC7 PPYR1 PRKG1 PRODH PTPLA protein phosphatase 2C homolog 7 pancreatic polypeptide receptor 1 protein kinase, cGMP-dependent, type I proline dehydrogenase (oxidase) 1 protein tyrosine phosphatase-like (proline instead of catalytic arginine), member A -1.6 2.5 1.8 2.9 1.7 4.64E-02 2.86E-02 3.54E-03 5.39E-10 7.56E-03 protein tyrosine phosphatase, receptor type, U pseudouridylate synthase 1 peroxidasin homolog (Drosophila)-like Uncharacterized protein 2.3 -1.6 3.8 1.5 1.53E-04 2.21E-02 1.26E-04 3.98E-02 PHACTR 2 PHYHD1 PIK3IP1 PIP5K1B Reference gene name platelet-derived growth factor receptor-like pyruvate dehydrogenase kinase, isozyme 4 6-phosphofructo-2-kinase/fructose-2,6biphosphatase 3 phosphatase and actin regulator 2 phytanoyl-CoA dioxygenase domain containing 1 phosphoinositide-3-kinase interacting protein 1 phosphatidylinositol-4-phosphate 5-kinase, type I, beta PLA2G15 phospholipase A2, group XV PLCD4 PTPRU PUS1 PXDNL Q2PP39_ MELGA ! 114 Table B.1 (cont’d) Gene ID Q2QF78_ MELGA Q90VX1_ MELGA Q90X50_ MELGA RAB3IP RASGEF 1B RBM20 REEP5 RELL1 RERGL RGS2 RGS7BP RNF20 RPH3AL RPIA Reference gene name Uncharacterized protein FC -3.7 FDR 4.72E-10 ATP synthase subunit alpha -1.8 6.72E-04 Uncharacterized protein -1.9 1.52E-03 RAB3A interacting protein (rabin3) RasGEF domain family, member 1B -2.2 2.7 2.54E-02 4.66E-03 RNA binding motif protein 20 receptor accessory protein 5 RELT-like 1 RERG/RAS-like regulator of G-protein signaling 2, 24kDa regulator of G-protein signaling 7 binding protein ring finger protein 20 rabphilin 3A-like (without C2 domains) ribose 5-phosphate isomerase A -1.9 1.8 1.6 1.8 -3.4 -3.0 1.5 2.1 1.5 4.59E-04 3.19E-03 2.51E-02 3.52E-02 2.13E-16 3.44E-02 3.84E-02 1.89E-02 4.10E-02 RPS23 RREB1 RRM2B RTN4RL1 S100A10 S100A6 S1PR3 SAMD11 SBK2 SCG5 SEC24D SEMA3C ribosomal protein S23 ras responsive element binding protein 1 ribonucleotide reductase M2 B (TP53 inducible) reticulon 4 receptor-like 1 S100 calcium binding protein A10 S100 calcium binding protein A6 sphingosine-1-phosphate receptor 3 sterile alpha motif domain containing 11 SH3-binding domain kinase family, member 2 secretogranin V (7B2 protein) SEC24 family, member D (S. cerevisiae) sema domain, immunoglobulin domain (Ig), short basic domain, secreted, (semaphoring) 3C sema domain, immunoglobulin domain (Ig), short basic domain, secreted, (semaphoring) 3D 1.5 -2.2 -1.7 6.8 1.5 1.7 2.3 -1.6 -2.7 2.4 1.8 1.5 4.86E-02 6.91E-06 1.12E-02 1.95E-02 4.10E-02 4.39E-02 4.45E-02 2.57E-02 2.68E-08 3.29E-03 8.53E-03 4.03E-02 1.9 5.14E-03 sema domain, immunoglobulin domain (Ig), short basic domain, secreted, (semaphoring) 3G -1.8 1.50E-03 SEMA3D SEMA3G ! 115 Table B.1 (cont’d) Gene ID SEMA6D Reference gene name sema domain, transmembrane domain ™, and cytoplasmic domain, ( emaphoring) 6D FC 1.7 FDR 1.00E-02 SERINC5 serine incorporator 5 SERPINE serpin peptidase inhibitor, clade E (nexin, 2 plasminogen activator inhibitor type 1), member 2 1.7 2.4 3.25E-02 1.92E-05 SERPINF serpin peptidase inhibitor, clade F (alpha-2 1 antiplasmin, pigment epithelium derived factor), member 1 SERPINF serpin peptidase inhibitor, clade F (alpha-2 2 antiplasmin, pigment epithelium derived factor), member 2 SERPINH serpin peptidase inhibitor, clade H (heat shock 1 protein 47), member 1, (collagen binding protein 1) SESN1 sestrin 1 SESN3 sestrin 3 SFRP2 secreted frizzled-related protein 2 SGSM2 small G protein signaling modulator 2 SIK2 salt-inducible kinase 2 SKA1 spindle and kinetochore associated complex subunit 1 SKIV2L2 superkiller viralicidic activity 2-like 2 (S. cerevisiae) SLC16A9 solute carrier family 16, member 9 (monocarboxylic acid transporter 9) SLC25A2 solute carrier family 25 (mitochondrial carrier; 4 phosphate carrier), member 24 SLC25A2 solute carrier family 38, member 2 9 SLC38A2 solute carrier family 43, member 2 SLC43A2 solute carrier family 43, member 2 SLC6A9 solute carrier family 6 (neurotransmitter transporter, glycine), member 9 SLC7A5 solute carrier family 7 (amino acid transporter light chain, L system), member 5 1.7 7.19E-03 2.8 1.89E-05 1.9 3.22E-04 -4.0 2.0 3.3 -2.2 -1.6 -2.5 9.29E-20 1.39E-03 4.75E-07 1.28E-04 1.78E-02 3.31E-02 1.6 2.54E-02 -2.6 8.38E-06 1.6 1.91E-02 -3.1 2.79E-08 1.6 -5.2 3.3 2.73E-02 1.46E-27 8.14E-03 2.5 1.17E-07 SLIT3 4.7 1.74E-06 SMPX ! slit homolog 3 (Drosophila) small muscle protein, x-linked 1.5 4.52E-02 116 Table B.1 (cont’d) Gene ID SPARC SPON2 SPP1 SPSB1 SQSTM1 SRPK3 SRPX ST8SIA2 STMN1 SYBU SYNGR3 SYNM TAGLN TARS TBC1D2 TC2N TFPI2 TGFB3 THRSP THY1 TIMP2 TLN1 TMED3 TMEM10 8 TMEM16 3 TMEM20 0A TMEM45 A TMEM9 ! Reference gene name secreted protein, acidic, cysteine-rich (osteonectin) spondin 2, extracellular matrix protein secreted phosphoprotein 1 splA/ryanodine receptor domain and SOCS box containing 1 sequestosome 1 SRSF protein kinase 3 sushi-repeat containing protein, X-linked ST8 alpha-N-acetyl-neuraminide alpha-2,8sialyltransferase 2 stathmin 1 syntabulin (syntaxin-interacting) synaptogyrin 3 synemin, intermediate filament protein transgelin threonyl-tRNA synthetase TBC1 domain family, member 2 tandem C2 domains, nuclear tissue factor pathway inhibitor 2 transforming growth factor, beta 3 thyroid hormone responsive Thy-1 cell surface antigen TIMP metallopeptidase inhibitor 2 talin 1 transmembrane emp24 protein transport domain containing 3 transmembrane protein 108 FC 1.8 FDR 8.04E-03 1.9 2.9 -1.6 2.67E-03 4.54E-03 1.11E-02 -1.7 2.2 2.1 -1.8 1.00E-02 5.10E-04 1.39E-03 1.14E-03 1.9 3.3 2.2 1.6 2.9 1.8 1.7 -2.1 1.9 1.6 -1.6 3.2 1.6 1.5 2.4 2.46E-02 2.30E-12 5.13E-03 2.40E-02 3.85E-09 1.19E-03 3.85E-02 2.54E-02 3.16E-02 6.63E-03 4.14E-02 3.57E-12 2.45E-02 3.73E-02 2.20E-04 1.9 7.83E-03 transmembrane protein 163 1.7 3.55E-02 transmembrane protein 200A 1.8 2.57E-02 transmembrane protein 45A 2.3 3.93E-03 transmembrane protein 9 -2.4 3.22E-04 117 Table B.1 (cont’d) Gene ID TMTC1 TNFAIP2 TRIM63 TUBA1C TUBB6 TULP1 TWIST2 UACA UBE2B UBXN7 USP2 USP35 VASH2 VASN VEGFA VGLL2 VIM WDFY1 WNT16 Reference gene name transmembrane and tetratricopeptide repeat containing 1 tumor necrosis factor, alpha-induced protein 2 tripartite motif containing 63 tubulin, alpha 1c tubulin, beta 6 tubby like protein 1 twist homolog 2 (Drosophila) uveal autoantigen with coiled-coil domains and ankyrin repeats ubiquitin-conjugating enzyme E2B UBX domain protein 7 ubiquitin specific peptidase 2 ubiquitin specific peptidase 35 vasohibin 2 vasorin vascular endothelial growth factor A vestigial like 2 (Drosophila) vimentin WD repeat and FYVE domain containing 1 wingless-type MMTV integration site family, member 16 FC -2.0 FDR 1.25E-03 -1.7 -1.6 2.0 2.2 -4.1 4.2 -2.3 1.53E-02 3.79E-02 6.33E-05 3.13E-05 1.47E-06 3.42E-03 8.30E-07 -1.6 -1.7 -2.3 -3.1 3.8 2.2 1.8 1.7 1.9 1.9 14.5 1.63E-02 7.05E-03 6.83E-07 4.70E-03 2.96E-19 3.95E-03 2.28E-03 1.14E-03 1.89E-03 5.96E-04 4.59E-04 WNT5B wingless-type MMTV integration site family, member 5B -1.6 3.16E-02 WNT9A wingless-type MMTV integration site family, member 9A -4.9 1.61E-02 XBP1 XPA X-box binding protein 1 xeroderma pigmentosum, complementation group A 2.3 2.2 9.97E-07 5.24E-06 xenotropic and polytropic retrovirus receptor 1 xylosyltransferase I tyrosyl-tRNA synthetase Yip1 domain family, member 7 zinc finger and BTB domain containing 16 zinc finger and BTB domain containing 44 -1.6 2.2 1.7 -1.6 -6.4 -1.8 1.53E-02 6.10E-03 5.73E-03 2.86E-02 9.35E-32 4.36E-03 XPR1 XYLT1 YARS YIPF7 ZBTB16 ZBTB44 ! 118 Table B.1 (cont’d) Gene ID ZNF516 unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown ! Reference gene name zinc finger protein 516 unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown 119 FC 2.5 -569.3 5.5 -8.3 -5.1 3.7 -9.8 -2.7 -2.9 -7.5 -3.5 -4.1 3.4 -2.5 -2.5 2.5 2.5 -2.8 2.6 2.1 -2.2 -6.2 -2.7 2.3 3.3 1.9 -2.0 7.2 4.7 1.9 -2.0 -1.9 1.9 3.5 FDR 1.36E-03 1.62E-55 1.68E-23 8.02E-20 2.91E-18 1.79E-11 2.08E-09 2.38E-09 2.06E-08 3.15E-08 2.42E-07 2.46E-07 4.76E-07 9.67E-07 1.18E-06 2.17E-06 2.61E-06 2.99E-06 3.58E-06 7.06E-06 1.36E-05 4.32E-05 4.35E-05 1.03E-04 1.49E-04 1.53E-04 2.51E-04 3.09E-04 3.09E-04 3.14E-04 4.88E-04 5.43E-04 7.29E-04 8.84E-04 Table B.1 (cont’d) Gene ID unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown ! Reference gene name unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown unknown 120 FC -1.9 -1.8 -12.5 1.8 -1.8 3.6 3.0 2.7 -1.7 5.4 1.7 4.2 -1.7 2.2 -1.7 2.1 -1.8 -1.7 1.6 1.7 1.7 -2.0 -1.8 -1.7 2.6 -1.6 -4.2 1.6 2.4 3.0 1.8 1.7 2.1 -1.5 FDR 1.01E-03 1.03E-03 1.44E-03 1.75E-03 1.98E-03 2.28E-03 2.42E-03 2.93E-03 3.94E-03 4.44E-03 4.48E-03 5.38E-03 7.20E-03 7.20E-03 9.52E-03 9.52E-03 9.74E-03 9.86E-03 1.01E-02 1.01E-02 1.20E-02 1.35E-02 1.59E-02 1.71E-02 1.93E-02 1.99E-02 2.05E-02 2.13E-02 2.18E-02 2.49E-02 2.51E-02 3.13E-02 3.44E-02 3.75E-02 Table B.1 (cont’d) Gene ID unknown unknown unknown unknown unknown unknown unknown unknown unknown Reference gene name unknown unknown unknown unknown unknown unknown unknown FC 1.5 1.6 -1.5 2.5 2.2 1.5 2.4 FDR 3.89E-02 3.95E-02 4.03E-02 4.03E-02 4.08E-02 4.18E-02 4.25E-02 unknown unknown 1.6 1.6 4.96E-02 4.88E-02 1 Fold change (FC) indicates ratio of expression of gene in PSE turkey relative to normal samples. Positive FC values show increased expression in PES relative to normal samples; negative FC values indicate decreased expression in PSE sample relative to normal. ! 121 REFERENCES ! 122 REFERENCES Aberle ED, Forrest JC, Gerrard DE, Mills EW. 2001. Principles of Meat Science, 4th ed. Iowa: Kendall/Hunt Publishing Company. Anthony NB. 1998. A review of genetic practices in poultry: Efforts to improve meat quality. Journal of Muscle Foods 9(1):25-33. Anders S, Huber W. 2010. Differential expression analysis for sequence count data. Genome Biology 11(10). Bacon WL, Nestor KE, Emmerson DA, Vasilatosyounken R, Long DW. 1993. Circulating IGF-I in plasma of growing male and female turkeys of medium and heavy weight lines. Domestic Animal Endocrinology 10(4):267-77. Baginsky S, Hennig L, Zimmermann P, Gruissem W. 2010. Gene expression analysis, proteomics, and network discovery. Plant Physiology 152(2):402-10. Barbut S. 1993. Color measurements for evaluating the pale soft exudative (PSE) occurrence in turkey meat. Food Research International 26(1):39-43. Barbut S. 1996. Estimates and detection of the PSE problem in young turkey breast meat. Canadian Journal of Animal Science 76(3):455-7. Barbut S. 1997a. Occurrence of pale soft exudative meat in mature turkey hens. British Poultry Science 38(1):74-7. Barbut S. 1997b. Problem of pale soft exudative meat in broiler chickens. British Poultry Science 38(4):355-8. Barbut S. 1998. Estimating the magnitude of the PSE problem in poultry. Journal of Muscle Foods 9(1):35-49. Barbut S. 2002. Poultry Products Processing: An industry guidlines, 10th ed. Boco Raton: CRC Press. Barbut S, Sosnicki AA, Lonergan SM, Knapp T, Ciobanu DC, Gatcliffe LJ, HuffLonergan E, Wilson EW. 2008. Progress in reducing the pale, soft and exudative (PSE) problem in pork and poultry meat. Meat Science 79(1):46-63. Barbut S, Zhang L, Marcone M. 2005. Effects of pale, normal, and dark chicken breast meat on microstructure, extractable proteins, and cooking of marinated fillets. Poultry Science 84(5):797-802. ! 123 Batlle N, Aristoy MC, Toldra F. 2000. Early postmortem detection of exudative pork meat based on nucleotide content. Journal of Food Science 65(3):413-6. Bianchi M, Fletcher DL. 2002. Effects of broiler breast meat thickness and background on color measurements. Poultry Science 81(11):1766-9. Bianchi M, Fletcher DL, Smith DP. 2005. Physical and functional properties of intact and ground pale broiler breast meat. Poultry Science 84(5):803-8. Bianchi M, Petracci M, Sirri F, Folegatti E, Franchini A, Meluzzi A. 2007. The influence of the season and market class of broiler chickens on breast meat quality traits. Poultry Science 86(5):959-63. Cassens RG. 2000. Historical perspectives and current aspects of pork meat quality in the USA. Food Chemistry 69(4):357-63. Chiang W, Allison CP, Linz JE, Strasburg GM. 2004. Identification of two alpha RYR alleles and characterization of alpha RYR transcript variants in turkey skeletal muscle. Gene 330:177-84. Chiang W, Booren A, Strasburg G. 2008. The effect of heat stress on thyroid hormone response and meat quality in turkeys of two genetic lines. Meat Science 80(3):615-22. Chiang W, Yoon HJ, Linz JE, Airey JA, Strasburg GM. 2007. Divergent mechanisms in generating molecular variations of alpha RYR and beta RYR in turkey skeletal muscle. Journal of Muscle Research and Cell Motility 28(6):343-54. Conlon MA, Kita K. 2002. Muscle protein synthesis rate is altered in response to a single injection of insulin-like growth factor-I in seven-day-old Leghorn chicks. Poultry Science 81(10). Connaughton S, Chowdhury F, Attia RR, Song SL, Zhang Y, Elam MB, Cook GA, Park EA. 2010. Regulation of pyruvate dehydrogenase kinase isoform 4 (PDK4) gene expression by glucocorticoids and insulin. Molecular and Cellular Endocrinology 315(1-2):159-67. Czerwinski SM, Cate JM, Francis G, Tomas F, Brocht DM, McMurtry JP. 1998. The effect of insulin-like growth factor-I (IGF-I) on protein turnover in the meat-type chicken (Gallus domesticus). Comparative Biochemistry and Physiology CPharmacology Toxicology & Endocrinology 119(1). Dalloul RA, Long JA, Zimin AV, Aslam L, Beal K, Blomberg L, Bouffard P, Burt DW, Crasta O, Crooijmans R, Cooper K, Coulombe RA, De S, Delany ME, Dodgson JB, Dong JJ, Evans C, Frederickson KM, Flicek P, Florea L, Folkerts O, Groenen MAM, Harkins TT, Herrero J, Hoffmann S, Megens HJ, Jiang A, ! 124 de Jong P, Kaiser P, Kim H, Kim KW, Kim S, Langenberger D, Lee MK, Lee T, Mane S, Marcais G, Marz M, McElroy AP, Modise T, Nefedov M, Notredame C, Paton IR, Payne WS, Pertea G, Prickett D, Puiu D, Qioa D, Raineri E, Ruffier M, Salzberg SL, Schatz MC, Scheuring C, Schmidt CJ, Schroeder S, Searle SMJ, Smith EJ, Smith J, Sonstegard TS, Stadler PF, Tafer H, Tu ZJ, Van Tassell CP, Vilella AJ, Williams KP, Yorke JA, Zhang LQ, Zhang HB, Zhang XJ, Zhang Y, Reed KM. 2010. Multi-Platform Next-Generation Sequencing of the Domestic Turkey (Meleagris gallopavo): Genome Assembly and Analysis. PLoS Biology 8(9). Donoghue P, Doran P, Wynne K, Pedersen K, Dunn MJ, Ohlendieck K. 2007. Proteomic profiling of chronic low-frequency stimulated fast muscle. Proteomics 7(18):3417-30. Dos Remedios CG, Chhabra D, Kekic M, Dedova IV, Tsubakihara M, Berry DA, Nosworthy NJ. 2003. Actin binding proteins: Regulation of cytoskeletal microfilaments. Physiological Reviews 83(2):433-73. Dransfield E, Sosnicki AA. 1999. Relationship between muscle growth and poultry meat quality. Poultry Science 78(5):743-6. Droval AA, Benassi VT, Rossa A, Prudencio SH, Paiao FG, Shimokomaki M. 2012. Consumer attitudes and preferences regarding pale, soft, and exudative broiler breast meat. Journal of Applied Poultry Research 21(3):502-7. Eadmusik S, Molette C, Fernandez X, Remignon H. 2011. Are one early muscle pH and one early temperature measurement sufficient to detect PSE breast meat in turkeys? British Poultry Science 52(2):177-88. Engel AG, Sine SM. 2005. Current understanding of congenital myasthenic syndromes. Current Opinion in Pharmacology 5(3):308-21. Fernandez X, Sante V, Baeza E, Lebihan-Duval E, Berri C, Remignon H, Babile R, Le Pottier G, Astruc T. 2002. Effects of the rate of muscle post mortem pH fall on the technological quality of turkey meat. British Poultry Science 43(2):24552. Fernandez X, Sante V, Baeza E, Lebihan-Duval E, Berri C, Remignon H, Babile R, Le Pottier G, Millet N, Berge P, Astruc T. 2001. Post mortem muscle metabolism and meat quality in three genetic types of turkey. British Poultry Science 42(4):462-9. Food and Agriculture Organization of the United Nations. [Internet]. Food and agricultural commodities production. [Accessd 2013 Mar 19]. Available from: http://faostat.fao.org/site/339/default.aspx ! 125 Franck M, Figwer P, Godfraind C, Poirel MT, Khazzaha A, Ruchoux MM. 2007. Could the pale, soft, and exudative condition be explained by distinctive histological characteristics? Journal of Animal Science 85(3):746-53. Fraqueza MJ, Cardoso AS, Ferreira MC, Barreto AS. 2006. Incidence of pectoralis major turkey muscles with light and dark color in a Portuguese slaughterhouse. Poultry Science 85(11):1992-2000. Fujii J, Otsu K, Zorzato F, Deleon S, Khanna VK, Weiler JE, Obrien PJ, Maclennan DH. 1991. Identification of a mutation in porcine ryanodine receptor associated with malignant hyperthermia. Science 253(5018):448-51. Garcia RG, de Freitas LW, Schwingel AW, Farias RM, Caldara FR, Gabriel AMA, Graciano JD, Komiyama CM, Paz I. 2010. Incidence and Physical Properties of PSE Chicken Meat in a Commercial Processing Plant. Brazilian Journal of Poultry Science 12(4):233-7. Golding-Myers JD, Showers CD, Shand PJ, Rosser BWC. 2010. Muscle fiber type and the occurrence of pale, soft and exudative pork. Journal of Muscle Foods 21(3):484-98. Goldman D, Brenner HR, Heinemann S. 1988. Acetylcholine-receptor alpha-subunit, beta-subunit, gamma-subunit and sigma-subunit messenger-RNA levels are regulated by muscle activity. Neuron 1(4):329-33. Gorsuch V, Alvarado CZ. 2010. Postrigor tumble marination strategies for improving color and water-holding capacity in normal and pale broiler breast fillets. Poultry Science 89(5):1002-8. Herbst EAF, Dunford ECE, Harris RA, Vandenboom R, LeBlanc PJ, Roy BD, Jeoung NH, Peters SJ. 2012. Role of pyruvate dehydrogenase kinase 4 in regulating PDH activation during acute muscle contraction. Applied Physiology Nutrition and Metabolism-Physiologie Appliquee Nutrition Et Metabolisme 37(1):48-52. Hitzemann R, Bottomly D, Darakjian P, Walter N, Iancu O, Searles R, Wilmot B, McWeeney S. 2013. Genes, behavior and next-generation RNA sequencing. Genes Brain and Behavior 12(1):1-12. Holness MJ, Kraus A, Harris RA, Sugden MC. 2000. Targeted upregulation of pyruvate dehydrogenase kinase (PDK)-4 in slow-twitch skeletal muscle underlies the stable modification of the regulatory characteristics of PDK induced by high-fat feeding. Diabetes 49(5):775-81. Jakubiec-Puka A, Ciechomska I, Morga J, Matusiak A. 1999. Contents of myosin heavy chains in denervated slow and fast rat leg muscles. Comparative ! 126 Biochemistry and Physiology B-Biochemistry & Molecular Biology 122(3):35562. Janisch S, Krischek C, Wicke M. 2011. Color values and other meat quality characteristics of breast muscles collected from 3 broiler genetic lines slaughtered at 2 ages. Poultry Science 90(8):1774-81. Jones JI, Clemmons DR. 1995. Insulin-like growth factors and their binding proteinsbiological actions. Endocrine Reviews 16(1):3-34. Kaprielian Z, Bandman E, Fambrough DM. 1991. Expression of Ca2+-ATPase isoforms in denervated, regenerating and dystrophic chicken skeletal muscle. Developmental Biology 144(1):199-211. Lahucky R, Baulain U, Henning M, Demo P, Krska P, Liptaj T. 2002. In vitro P-31 NMR studies on biopsy skeletal muscle samples compared with meat quality of normal and heterozygous malignant hyperthermia pigs. Meat Science 61(3):233-41. Lan J, Lei MG, Zhang YB, Wang JH, Feng XT, Xu DQ, Gui JF, Xiong YZ. 2009. Characterization of the porcine differentially expressed PDK4 gene and association with meat quality. Molecular Biology Reports 36(7):2003-10. Laville E, Sayd T, Sante-Lhoutellier V, Morzel M, Labas R, Franck M, Chambon C, Monin G. 2005. Characterisation of PSE zones in semimembranosus pig muscle. Meat Science 70(1):167-72. Le Bihan-Duval E, Berri C, Baeza E, Sante V, Astruc T, Remignon H, Le Pottier G, Bentley J, Beaumont C, Fernandezb X. 2003. Genetic parameters of meat technological quality traits in a grand-parental commercial line of turkey. Genetics Selection Evolution 35(6):623-35. Lilburn MS, Nestor KE. 1991. Body-weight and carcass development in different lines of turkeys. Poultry Science 70(11):2223-31. Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression data using realtime quantitative PCR and the 2(T)(-Delta Delta C) method. Methods 25(4):402-8. Lytton J, Westlin M, Burk SE, Shull GE, Maclennan DH. 1992. Functional comparisons between isoforms of the sarcoplasmic or endoplasmic-reticulum family of calcium pumps. Journal of Biological Chemistry 267(20):14483-9. Malila Y, Carr KM, Ernst CW, Velleman SG, Reed KM, Strasburg GM. 2013. Deep transcriptome sequencing reveals difference in global gene expression between normal and pale, soft and exudative turkey meat ! 127 (in preparation) Malila Y, Tempelman RJ, Sporer KRB, Ernst CW, Velleman SG, Reed KM, Strasburg GM. 2013. Differential gene expression between normal and pale, soft, and exudative turkey meat. Poultry science 92(6):1621-33. Manning DR, Stull JT. 1982. Myosin light chain phosphorylation-dephosphorylation in mammalian skeletal muscle. American Journal of Physiology 242(3):C234C41. Marchi DF, Oba A, Ziober IL, Soares AL, Ida EI, Shimokomaki M. 2009. Development of a gas chamber for detecting broiler chicken halothane sensitivity and PSE (Pale, Soft, Exudative) meat formation. Brazilian Archives of Biology and Technology 52:189-94. Marioni JC, Mason CE, Mane SM, Stephens M, Gilad Y. 2008. RNA-seq: An assessment of technical reproducibility and comparison with gene expression arrays. Genome Research 18(9):1509-17. McClung JM, Thompson RW, Lowe LL, Carson JA. 2004. RhoA expression during recovery from skeletal muscle disuse. Journal of Applied Physiology 96(4):1341-8. McCurdy RD, Barbut S, Quinton M. 1996. Seasonal effect on pale soft exudative (PSE) occurrence in young turkey breast meat. Food Research International 29(3-4):363-6. McKee SR, Sams AR. 1997. The effect of seasonal heat stress on rigor development and the incidence of pale, exudative turkey meat. Poultry Science 76(11):1616-20. McMurtry JP, Francis GL, Upton Z. 1997. Insulin-like growth factors in poultry. Domestic Animal Endocrinology 14(4):199-229. Mickelson JR, Gallant EM, Litterer LA, Johnson KM, Rempel WE, Louis CF. 1988a. Abnormal sarcoplasmic reticulum ryanodine receptor in malignant hyperthermia. Journal of Biological Chemistry 263(19):9310-5. Mickelson JR, Johnson KM, Litterer LA, Louis CF. 1988b. Abnormal ryanodine receptor properties correlate with altered calcium release of sarcoplasmicreticulum from malignant hyperthermia susceptible pigs. Biophysical Journal 53(2):A336-A. Mickelson JR, Louis CF. 1996. Malignant hyperthermia: Excitation-contraction coupling, Ca2+ release channel, and cell Ca2+ regulation defects. Physiological Reviews 76(2):537-92. ! 128 Mickelson JR, Ross JA, Reed BK, Louis CF. 1986. Enhanced Ca2+-induced calcium release by isolated sarcoplasmic reticulum vesicles from malignant hyperthermia susceptible pig muscle. Biochimica Et Biophysica Acta 862(2):318-28. Miri A, Talmant A, Renou JP, Monin G. 1992. P-31 NMR-study of post-mortem changes in pig muscle. Meat Science 31(2):165-73. Moesgaard B, Quistorff B, Christensen VG, Therkelsen I, Jorgensen PF. 1995. Differences of postmortem ATP turnover in skeletal muscle of normal and heterozygote malignant hyperthermia pigs-comparison of P31-NMR and analytical biochemical measurements. Meat Science 39(1):43-57. Nestor K. 1977a. The use of a paired mating system for the maintenance of experimental populations of turkeys. Poultry Science 56:60-5. Nestor KE. 1977b. Genetics of growth and reproduction in turkey. 5. selection for increased body-weight alone and in combination with increased egg production. Poultry Science 56(1):337-47. Nestor KE, McCartney MG, Harvey WR. 1967. Genetics of growth and reproduction in turkey. 1. genetic and non-genetic variation in body weight and body measurements. Poultry Science 46(6):1374-84. Nierobisz LS, Sporer KRB, Strasburg GM, Reed KM, Velleman SG, Ashwell CM, Felts JV, Mozdziak PE. 2012. Differential expression of genes characterizing myofibre phenotype. Animal Genetics 43(3):298-308. Obi T, Matsumoto M, Miyazaki K, Kitsutaka K, Tamaki M, Takase K, Miyamoto A, Oka T, Kawamoto Y, Nakada T. 2010. Skeletal ryanodine receptor 1heterozygous PSE (pale, soft and exudative) meat contains a higher concentration of myoglobin than genetically normal PSE meat in pigs. The Asian-Australian Association of Animal Production Socieities 23(9):6. Oda SHI, Nepomuceno AL, Ledur MC, de Oliveira MCN, Marin SRR, Ida EI, Shimokomaki M. 2009. Quantitative Differential Expression of Alpha and Beta Ryanodine Receptor Genes in PSE (Pale, Soft, Exudative) Meat from Two Chicken Lines: Broiler and Layer. Brazilian Archives of Biology and Technology 52(6):1519-25. Odermatt A, Barton K, Khanna VK, Mathieu J, Escolar D, Kuntzer T, Karpati G, MacLennan DH. 2000. The mutation of Pro(789) to Leu reduces the activity of the fast-twitch skeletal muscle sarco(endo)plasmic reticulum Ca2+ ATPase (SERCA1) and is associated with Brody disease. Human Genetics 106(5):48291. ! 129 Otsu K, Nishida K, Kimura Y, Kuzuya T, Hori M, Kamada T, Tada M. 1994. The Point mutation Arg615 -> Cys in the Ca2+ release channel of skeletal sarcoplasmic reticulum is responsible for hypersensitivity to caffeine and halothane in malignant hyperthermia. Journal of Biological Chemistry 269(13):9413-5. Owens CM, Alvarado CZ, Sams AR. 2009. Research developments in pale, soft, and exudative turkey meat in North America. Poultry Science 88(7):1513-7. Owens CM, Hirschler EM, McKee SR, Martinez-Dawson R, Sams AR. 2000. The characterization and incidence of pale, soft, exudative turkey meat in a commercial plant. Poultry Science 79(4):553-8. Park I, Han C, Jin S, Lee B, Choi H, Kwon JT, Kim D, Kim J, Lifirsu E, Park WJ, Park ZY, Kim DH, Cho C. 2011. Myosin regulatory light chains are required to maintain the stability of myosin II and cellular integrity. Biochemical Journal 434:171-80. Pedersen SE, Cohen JB. 1990. D-tubocurarine binding-sites are located at alphagamma and acetylcholine receptor. Proceedings of the National Academy of Sciences of the United States of America 87(7):2785-9. Petracci M, Betti M, Bianchi M, Cavani C. 2004. Color variation and characterization of broiler breast meat during processing in Italy. Poultry Science 83(12):208692. Petracci M, Bianchi M, Cavani C. 2009. The European perspective on pale, soft, exudative conditions in poultry. Poultry Science 88(7):1518-23. Petracci M, Cavani C. 2012. Muscle growth and poultry meat quality issues. Nutrients 4(1):1-12. Petracci M, Fletcher DL. 2002. Broiler skin and meat color changes during storage. Poultry Science 81(10):1589-97. Pette D, Staron RS. 2001. Transitions of muscle fiber phenotypic profiles. Histochemistry and Cell Biology 115(5):359-72. Pette, D, Vrbová G. 1999. What does chronic electrical stimulation teach us about muscle plasticity? Muscle Nerve 22:666-677. Pietrzak M, Greaser ML, Sosnicki AA. 1997. Effect of rapid rigor mortis processes on protein functionality in pectoralis major muscle of domestic turkeys. Journal of Animal Science 75(8):2106-16. Popov KM. 1997. Regulation of mammalian pyruvate dehydrogenase kinase. Febs Letters 419(2-3). ! 130 Preiss T, Baron-Benhamou J, Ansorge W, Hentze MW. 2003. Homodirectional changes in transcriptome composition and mRNA translation induced by rapamycin and heat shock. Nature Structural Biology 10(12):1039-47. Randle PJ, Garland PB, Newsholme EA, Hales CN. 1963. Glucose fatty acid cycle – Its role in insulin sensitivity and metabolic disturbances of diabetes mellitus. Lancet 1(728):785-&. Randle PJ, Priestman DA, Mistry SC, Halsall A. 1994. Glucose fatty acid interactions and the regulation of glucose disposal. Journal of Cellular Biochemistry 55:111. Rasband WS. 2012. ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, http://imagej.nih.gov/ij/ Reed KM, Mendoza KM, Juneja B, Fahrenkrug SC, Velleman S, Chiang W, Strasburg GM. 2008. Characterization of expressed sequence tags from turkey skeletal muscle. Animal Genetics 39(6):635-44. Richards MP, Poch SM, McMurtry JP. 2005. Expression of insulin-like growth factor system genes in liver and brain tissue during embryonic and post-hatch development of the turkey. Comparative Biochemistry and Physiology aMolecular & Integrative Physiology 141(1):76-86. Ritchie ME, Silver J, Oshlack A, Holmes M, Diyagama D, Holloway A, Smyth GK. 2007. A comparison of background correction methods for two-colour microarrays. Bioinformatics 23(20):2700-7. Roberts A, Trapnell C, Donaghey J, Rinn J, Pachter L. 2011. Improving RNA-Seq expression estimates by correcting for fragment bias. Genome Biology 12(3). Roh SW, Abell GCJ, Kim KH, Nam YD, Bae JW. 2010. Comparing microarrays and next-generation sequencing technologies for microbial ecology research. Trends in Biotechnology 28(6):291-9. Rosser BWC, Wick M, Waldbillig DM, Bandman E. 1996. Heterogeneity of myosin heavy-chain expression in fast twitch fiber types of mature avian pectoralis muscle. Biochemistry and Cell Biology-Biochimie Et Biologie Cellulaire 74(5):715-28. Ryu YC, Kim BC. 2006. Comparison of histochemical characteristics in various pork groups categorized by postmortem metabolic rate and pork quality. Journal of Animal Science 84(4):894-901. ! 131 Sacchetto R, Testoni S, Gentile A, Damiani E, Rossi M, Liguori R, Droegemueller C, Mascarello F. 2009. A defective SERCA1 protein is responsible for congenital pseudomyotonia in Chianina cattle. American Journal of Pathology 174(2):565-73. Samuel DD, Billard L, Pringle D, Wicker L. 2012. Influence of growth rate on occurrences of pale muscle in broilers. Journal of the Science of Food and Agriculture 92(1):78-83. Schmidt A, Hall MN. 1998. Signaling to the actin cytoskeleton. Annual Review of Cell and Developmental Biology 14:305-38. Shen QW, Underwood KR, Means WJ, McCormick RJ, Du M. 2007. The halothane gene, energy metabolism, adenosine monophosphate-activated protein kinase, and glycolysis in postmortem pig longissimus dorsi muscle. Journal of Animal Science 85(4):1054-61. Sheterline P, Sparrow JC. 1994. Actin. Protein Profile 1(1):1-121. Shi LM, Jones WD, Jensen RV, Harris SC, Perkins RG, Goodsaid FM, Guo L, Croner LJ, Boysen C, Fang H, Qian F, Amur S, Bao WJ, Barbacioru CC, Bertholet V, Cao XM, Chu TM, Collins PJ, Fan XH, Frueh FW, Fuscoe JC, Guo X, Han J, Herman D, Hong HX, Kawasaki ES, Li QZ, Luo YL, Ma YQ, Mei N, Peterson RL, Puri RK, Shippy R, Su ZQ, Sun YA, Sun HM, Thorn B, Turpaz Y, Wang C, Wang SJ, Warrington JA, Willey JC, Wu J, Xie Q, Zhang L, Zhong S, Wolfinger RD, Tong WD. 2008. The balance of reproducibility, sensitivity, and specificity of lists of differentially expressed genes in microarray studies. Bmc Bioinformatics 9. Sia SK, Li MX, Spyracopoulos L, Gagne SM, Liu W, Putkey JA, Sykes BD. 1997. Structure of cardiac muscle troponin C unexpectedly reveals a closed regulatory domain. Journal of Biological Chemistry 272(29):18216-21. Sibut V, Hennequet-Antier C, Le Bihan-Duval E, Marthey S, Duclos MJ, Berri C. 2011. Identification of differentially expressed genes in chickens differing in muscle glycogen content and meat quality. Bmc Genomics 12:13. Smyth GK. 2005. Limma: linear models for microarray data. In: Gentleman R, Carey V, Dudoit S, Irizarry R, Huber W, editors. Bioinformatics and Computational Biology Solutions using R and Bioconductor. New York: Springer. p. 397-420. Sohn RH, Goldschmidtclermont PJ. 1994. Profilin-at the crossroads of signaltransduction and the actin cytoskeleton. Bioessays 16(7):465-72. ! 132 Solomon MB, Van Laack R, Eastridge JS. 1998. Biophysical basis of pale, soft, exudative (PSE) pork and poultry muscle: A review. Journal of Muscle Foods 9(1):1-11. Sporer KRB, Chiang W, Tempelman RJ, Ernst CW, Reed KM, Velleman SG, Strasburg GM. 2011a. Characterization of a 6K oligonucleotide turkey skeletal muscle microarray. Animal Genetics 42(1):75-82. Sporer KRB, Tempelman RJ, Ernst CW, Reed KM, Velleman SG, Strasburg GM. 2011b. Transcriptional profiling identifies differentially expressed genes in developing turkey skeletal muscle. BMC Genomics 12. doi 14310.1186/14712164-12-143. Sporer KRB, Zhou HR, Linz JE, Booren AM, Strasburg GM. 2012. Differential expression of calcium-regulating genes in heat-stressed turkey breast muscle is associated with meat quality. Poultry Science 91(6):1418-24. Stromer MH. 1998. The cytoskeleton in skeletal, cardiac and smooth muscle cells. Histology and Histopathology 13(1):283-91. Stull JT, Kamm KE, Vandenboom R. 2011. Myosin light chain kinase and the role of myosin light chain phosphorylation in skeletal muscle. Archives of Biochemistry and Biophysics 510(2):120-8. Sugden MC, Bulmer K, Holness MJ. 2001. Fuel-sensing mechanisms integrating lipid and carbohydrate utilization. Biochemical Society Transactions 29:272-8. Sugden MC, Holness MJ. 2003. Recent advances in mechanisms regulating glucose oxidation at the level of the pyruvate dehydrogenase complex by PDKs. American Journal of Physiology-Endocrinology and Metabolism 284(5):E855E62. Tomas FM, Pym RA, McMurtry JP, Francis GL. 1998. Insulin-like growth factor (IGF)I but not IGF-II promotes lean growth and feed efficiency in broiler chickens. General and Comparative Endocrinology 110(3). Tonge DP, Jones SW, Bardsley RG, Parr T. 2010. Characterisation of the sarcomeric myosin heavy chain multigene family in the laboratory guinea pig. Bmc Molecular Biology 11. Trapnell C, Pachter L, Salzberg SL. 2009. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25(9):1105-11. Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L. 2010. Transcript assembly and quantification by RNA- ! 133 Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nature Biotechnology 28(5):511-U174. US Department of Agriculture. [Internet]. USDA Economics, Statistics and Market Information System: poultry yearbook and red meat yearbook. [Accessed 2013 April 17]. Available from: http://usda.mannlib.cornell.edu/ MannUsda/homepage.do Updike MS, Zerby HN, Sawdy JC, Lilburn MS, Kaletunc G, Wick MP. 2005. Turkey breast meat functionality differences among turkeys selected for body weight and/or breast yield. Meat Science 71(4):706-12. Vanderstoep.J, Richards JF. 1974. Postmortem glycolytic and physical changes in turkey breast muscle. Canadian Institute of Food Science and Technology Journal-Journal De L Institut Canadien De Science Et Technologie Alimentaires 7(2):120-4. Vanhoof J. 1979. Influence of ante-mortem and peri-mortem factors on biochemical and physical characteristics of turkey breast muscle. Veterinary Quarterly 1(1):29-36. Wang LJ, Byrem TM, Zarosley J, Booren AM, Strasburg GM. 1999. Skeletal muscle calcium channel ryanodine binding activity in genetically unimproved and commercial turkey populations. Poultry Science 78(5):792-7. Wang Z, Gerstein M, Snyder M. 2009. RNA-Seq: a revolutionary tool for transcriptomics. Nature Reviews Genetics 10(1):57-63. Warner RD, Kauffman RG, Greaser ML. 1997. Muscle protein changes post mortem in relation to pork quality traits. Meat Science 45(3):339-52. Wells L, Edwards KA, Bernstein SI. 1996. Myosin heavy chain isoforms regulate muscle function but not myofibril assembly. Embo Journal 15(17):4454-9. Wilhelm AE, Maganhini MB, Hernandez-Blazquez FJ, Ida EI, Shimokomaki M. 2010. Protease activity and the ultrastructure of broiler chicken PSE (pale, soft, exudative) meat. Food Chemistry 119(3):1201-4. Wilkins LJ, Brown SN, Phillips AJ, Warriss PD. 2000. Variation in the colour of broiler breast fillets in the UK. British Poultry Science 41(3):308-12. Woelfel RL, Owens CM, Hirschler EM, Martinez-Dawson R, Sams AR. 2002. The characterization and incidence of pale, soft, and exudative broiler meat in a commercial processing plant. Poultry Science 81(4):579-84. ! 134 Wu PF, Blair PV, Sato J, Jaskiewicz J, Popov KM, Harris RA. 2000. Starvation increases the amount of pyruvate dehydrogenase kinase in several mammalian tissues. Archives of Biochemistry and Biophysics 381(1):1-7. Wynn RM, Kato M, ChuangO JL, Tso SC, Li J, Chuang DT. 2008. Pyruvate dehydrogenase kinase-4 structures reveal a metastable open conformation fostering robust core-free basal activity. Journal of Biological Chemistry 283(37):25305-15. Zhang L, Barbut S. 2005. Effects of regular and modified starches on cooked pale, soft, and exudative; normal; and dry, firm, and dark breast meat batters. Poultry Science 84(5):789-96. Zhi G, Ryder JW, Huang J, Ding PG, Chen Y, Zhao YM, Kamm KE, Stull JT. 2005. Myosin light chain kinase and myosin phosphorylation effect frequencydependent potentiation of skeletal muscle contraction. Proceedings of the National Academy of Sciences of the United States of America 102(48):1751924. Zhu XS, Xu XL, Min HH, Zhou GH. 2012. Occurrence and Characterization of Pale, Soft, Exudative-Like Broiler Muscle Commercially Produced in China. Journal of Integrative Agriculture 11(8):1384-90. Ziober IL, Paiao FG, Marchi DF, Coutinho LL, Binneck E, Nepomuceno AL, Shimokomaki M. 2010. Heat and chemical stress modulate the expression of the alpha-RYR gene in broiler chickens. Genetics and Molecular Research 9(2):1258-66. ! 135