Development and application of molecular biomarkers for characterizing Caribbean Yellow Band Disease in Orbicella faveolata

Coral collections

Coral samples were collected in Oct 2009 and Oct 2013 from La Parguera, Puerto Rico. Specimens were collected under the General Collection Permit of the Department of Marine Sciences, University of Puerto Rico, Mayaguez. Diseased and healthy colonies of Orbicella faveolata were sampled from depth range of 9 to 12 m. At the time of collection, all samples were individually labeled, placed in plastic bags, and immediately transported back to the laboratory where they were placed on a seawater table and immediately processed. Five colonies with no visible evidence of disease were used as representative healthy tissue controls. Five other colonies with visible CYBD lesions were sampled as representative of the diseased condition. In addition, tissues were also sampled in the transition zone (Weil, Cróquer & Urreiztieta, 2009b), 2–4 cm in front of the visible lesion border on the same five diseased coral colonies. Diseased tissue (visible lesion) was identified and then separated from colony sample by chisel. Transition tissue was also taken from the same diseased colony. Transition tissues were sampled approximately 2–4 cm in front of the advancing visually identified transition/lesion border (see Fig. 1). Healthy colonies which had no visible signs of CYBD were also collected. After each sample was isolated, approximately 2 to 3 cm² of tissue was immersed in 5 mls of Trizol followed by immediate homogenization by vortexing. Total RNA was extracted following Trizol protocol (Invitrogen, Carlsbad, CA, USA) with the additional use of 2 ml phase-lock gels (5’Prime, Gaithersburg, MD) to aid in the recovery of the aqueous phase. RNA concentrations were estimated by absorbance readings at 260 nm (Biophotometer; Eppendorf, Hauppauge, NY). The integrity of the total RNA was confirmed by electrophoresis of an aliquot of each sample on a 1% formaldehyde agarose gel (Sambrook, Firtsch & Maniatia , 2001). Total RNA from each sample was purified by DNase I digestion followed by phenol/CHCl₃ extraction (Message Clean; GenHunter, Nashville, TN, USA). Messenger RNA (mRNA) was isolated (Oligotex; Qiagen, Valencia, CA, USA) from 100 µg of DNase I treated total RNA. Reverse transcription of 1 µg of mRNAs followed manufacturer’s protocol (cDNA synthesis; Invitrogen, Carlsbad, CA, USA) which also included the use of both oligo-dT and random hexamer primers. The RT-reaction conditions were modified to 1 h at 37 °C, followed by 1 min at each temperature between 42 and 50 °C to maximize the number of full-length and partial transcripts copied (Pastorian, Hawel & Byus, 2000).

Isolating RDA fragments

Hubank & Schatz (1999) provided the framework for the RDA protocol including RDA primer sequences (Table 1), with modifications to the amount of starting material (Edman et al., 1997) and the elimination of mung bean nuclease (Pastorian, Hawel & Byus, 2000). Double stranded cDNA from each treatment was digested with DpnII (New England Biolabs, Ipswich, MA, USA). Digested products were spin column purified to eliminate fragments smaller than 100 bp in length (QIAquick PCR purification; Qiagen, Valencia, CA, USA). Primers R12 and R24 were ligated onto the digested cDNAs at 15 °C overnight in 60 µl reactions. To generate sufficient quantities of the required amplicons necessary for downstream protocols, replicate PCR reactions were performed using 2.5 µl of ligated cDNAs that were amplified for 20 cycles. Prior to amplification, there was a 3 min incubation at 72 °C. Subsequently, primer was added and allowed to incubate for an additional 5 min. Once the entire 8 min heating had elapsed, Taq DNA polymerase was added and amplification proceeded for 20 cycles of 45 s at 95 °C followed by 3 min at 72 °C and then concluding for a final 10 min extension at 72 °C.

Gel electrophoresis confirmed the size distribution within each amplicon. Twenty replicate PCR reactions were pooled and then precipitated by isopropanol to concentrate. Each amplicon was quantified by UV spectrophotometry and subsequently diluted with TE buffer to a final concentration of 0.5 µg/µl. Amplicons that were digested with DpnII (New England Biolabs, Ipswich, MA, USA) to remove R24 primers became the cut-drivers which were used in downstream reactions. Two rounds of hybridization were performed in this investigation. With the assistance of J12 primer, the J24 primer was ligated onto cut-driver and used as the tester amplicon in the first round of hybridization. Tester populations in the second round of hybridizations used the N24/N12 primers. Ligation conditions were always the same throughout the investigation even though the use of J24/J12 or N24/N12 primers depended on the round of hybridization. The ratio of tester/driver in round one was 1:100, whereas in round 2 it was 1:800. The first round of hybridization combined 50 ng of J24-ligated tester to 5 µg of cut-driver. The second round of subtraction/hybridization combined 6.25 ng of N24-ligated hybridization product from the first round of hybridization and 5 µg of cut-driver. Each sample population (healthy or diseased) was used as a tester in one series of hybridizations and as the driver in the other series of hybridizations. After combining testers and drivers in the desired ratio, pooled samples were extracted with Phenol/Chloroform/Isoamyl alcohol and then precipitated in 30 µl of 10 M ammonium acetate and 250 µl ethanol at −70 °C for one hour. After centrifugation at 14,000 rpm for 15 min at 4 °C, resulting pellets were washed twice in 70% ethanol and allowed to air dry. Each pellet was resuspended in 4 µl EEx3 (30 mM EPPS, pH 8.0 at 20 °C, 3 mM EDTA) buffer by pipetting repeatedly for 2 min then warmed to 37 °C for 5 min, vortexed, and then briefly centrifuged. Samples were then overlayed with 35 µl mineral oil and then heated to 95 °C for 5 min to denature. Afterwards they were allowed to cool to 67 °C and 1 µl 5M NaCl was added directly into the DNA, and the samples were incubated overnight at 67 °C. After the overnight incubation, mineral oil was removed and samples were diluted in 95 µl TE. Five microliters of a hybridization product were used in downstream PCR reactions. The hybridization products were amplified by PCR beginning with a 3 min incubation at 72 °C followed by the addition of Taq DNA polymerase. After 5 additional minutes at 72 °C, primer was added (either J24 or N24 primer depending on which round of hybridization had been performed). Amplification conditions consisted of 45 s at 95 °C followed by 3 min at 70 °C (J24) or 72 °C (N24) for a total of 27 cycles and then concluded with a 10 min extension at 72 °C. To eliminate the need of mung bean nuclease, a modified PCR reaction was employed which involved taking a 10 µl aliquot out of the PCR reaction after 7 cycles and placing it in a new PCR reaction with all reagents and continuing for an additional 20 cycles (Pastorian, Hawel & Byus, 2000). Amplified RDA products were cloned (TopoTA, Invitrogen, Carlsbad, CA, USA) and then sequenced using M13 (forward and reverse) primers (Nevada Genomics Center, University of Reno, NV, USA). Ninety-five sequences from CYBD infected tissues were cloned and sequenced. Another 190 sequences were sequenced from RDA products representing healthy tissues.

Sequence analysis

A contig assembly program (CAP3) (http://doua.prabi.fr/software/cap3) was used to determine how many unique sequences were represented within the total number of cloned RDA products. Sequences were compared against the nr and EST databases at NCBI (National Center for Biotechnology Information) database (http://www.ncbi.nlm.nih.gov) using the BLASTX algorithm with default parameters. CAP3 analysis indicated there were 9 contigs and 53 singletons.

Virtual northern dot blot screen

Two microliter aliquots of individually amplified RDA products were blotted onto replicate positively charged nylon membranes. Amplicons were amplified by PCR using a DIG-labeled nucleotide (Roche, Indianapolis, IN, USA). The presence/abundance of individual RDA products was detected by chemiluminescence and quantified by densitometry (Morgan et al., 2012). One of the replicate membranes of RDA products was probed with DIG-labeled amplicon from the healthy tissues, while the other replicate membrane was probed with DIG-labeled amplicon from the diseased tissues. DIG-labeled RDA products with differences in signal intensities between the identical membranes were identified and selected as candidate GOIs for subsequent qPCR analyses.

Selection of GOIs and primer creation

Virtual Northerns (Franz, Bruchhaus & Roeder, 1999) and BLAST analyses identified 14 potential genes of interest (GOIs) that were evaluated as candidates as a qPCR control gene or a gene which is differentially expressed amongst the tissues (Table 2). Gene specific forward and reverse primers were created for the GOIs using Primer3 (http://frodo.wi.mit.edu/primer3/) (Table 3). Primer efficiency was initially screened by amplifying a target cDNA over a range of T_m’s to determine optimal annealing (Mastercycler gradient; Eppendorf, Hauppauge, NY) and then compared predicted amplicon size to observed amplicon size on a 2% TBE agarose gel. Five GOIs were chosen for further analyses (Table 4). Poly-A binding protein (PABP) was chosen as the reference gene because it has previously demonstrated a consist pattern of expression as a qPCR reference gene (Rodriquiez-Lanetty et al., 2008). Four of the five genes (EGR, Ubiquitin Ligase, Superoxide dismutase, and PABP) were representative of the coral host while the fifth gene (Cytochrome Oxidase subunit 1) was representative of zooxanthellae. Melt-curve analysis was performed to determine specificity of priming. Individual amplicons were also extracted from agarose gels (Qiaquick gel extraction kit, Qiagen, Valencia, CA, USA) and sequenced to confirm amplification of the intended target.

Quantitative real-time PCR

Quantitative Real Time PCR (qPCR) assays were performed in four replicate 25 µL reactions. The components within each reaction were: 12.5 µL (2X) PowerSYBR^® Green (Applied Biosystems, Carlsbad, CA), 2.5 µL of forward and reverse primer (10 µM each), 2.5 µL of 3-fold diluted cDNA from reverse-transcription reaction, 0.25 µL Taq DNA polymerase (BioLabs, Ipswich MA), and 4.75 µL DI water. All qPCR reactions were performed on a StepOne™ Real Time PCR machine (Applied Biosystems, Carlsbad, CA). Reaction conditions involved heating at 95 °C for 15 s, annealing for 54 °C or 56 °C for 20 s, and then elongating at 72 °C for 20 s. All reactions were monitored by a melt curve analysis to ensure specific amplification and absence of primer dimerization. Reactions were ramped from 60 °C to 95 °C at a rate of 0.3 °C s⁻¹.

Expression analysis

Replicate Cq values for each GOI were averaged to determine ΔCq and ΔΔCq for each sample within a health condition. All qPCR ΔCq and ΔΔCq values are based on the consistent expression of the PABP across all samples in this study. The ΔΔCq method was used to determine the differences between targeted GOIs and a single reference gene (Bustin et al., 2009). One-way ANOVA was used on the ΔΔCq data to identify significant differences in the expression of an individual GOI within tissues representing different stages of the disease (i.e., diseased, transition, healthy). Similarities in variance were determined by Levene’s Test of Equality of Error Variances. If the variance between disease stages was similar, then the Student-Neuman-Keuls (SNK) posthoc test was performed to determine which group(s) were significantly different from the rest. If variance between disease stages was different, then Tamhane’s T2 posthoc test was applied since One-way ANOVA is generally insensitive to heterosedcasticity.

Functional significance of GOI’s

Early growth response

Preliminary BLASTX results suggest that RDA product H1D1-64 is an early growth response (EGR) which is known as a transcription factor (Barshis et al., 2013). This gene is known to be associated with the initiation of immune responses (McMahon & Monroe, 1996), mitogenesis and cell growth (Sukhatme, 1990), and tumor suppression (Huang et al., 1997). Corals are known to have a complex repertoire of immune responses for how corals respond to a pathogen (Sutherland, Porter & Torres, 2004; Miller et al., 2007; Schwarz et al., 2008; Kvennefors et al., 2010; Shinzato et al., 2011; Poole & Weis, 2014). Coral immune mechanisms include coral wound healing, hemocytosis, phagocytosis, encapsulation, and basic immunological memory to fight off pathogens (Sutherland, Porter & Torres, 2004; Palmer & Traylor-Knowles, 2012). EGR has also previously demonstrated significant expression in corals responding to heat stress (Barshis et al., 2013) which is a well-recognized stressor that can induce bleaching. There is a growing body of evidence of linkages between bleaching and immune responses in cnidarians (Weis, 2008; Kvennefors et al., 2010; Detournay et al., 2012; Pinzon et al., 2014; Pratte & Richardson, 2014). Elevated EGR expression in the transition tissues (see Fig. 3) may represent an important signal associated with a preliminary stage of infection in nearby tissues, but further studies will need to be conducted to clarify the role of this protein in the progression of CYBD.

Ubiquitin ligase

Ubiquitin ligase in this study is representative of RING-type E3 ligases. Ubiquitination directs many cellular functions including protein degradation (Komander & Rape, 2012) and regulating a variety of cellular processes including vesicle trafficking (Hsu et al., 2014), cell cycle control (Skaar, Pagan & Pagano, 2013), and immune responses (Jiang & Chen, 2012). Results herein reveal ubiquitin ligase is initially up-regulated in transition tissues, but peaks with highest expression in visible lesion tissues (see Fig. 4). Elevated expression of ubiquitin ligase in diseased tissues suggests this profile is representative of a later stage in the development of the disease. Elevated expression of ubiquitin ligase in diseased tissues is particularly interesting for a couple of reasons. Microbes are known to highjack host ubiquitin pathways in order to manipulate host signaling to facilitate bacterial infection and proliferation (Zhou & Zhu, 2015), and Closek et al. (2014) identified CYBD diseased tissues having 2-3 times greater bacterial diversity compared to healthy tissues while the transition tissues actually had the highest species richness. Ubiquitination is also known to be associated with lysosomal degradation of plasma membrane proteins (Komander & Rape, 2012), therefore the expression profile of ubiquitin ligase in this study is also consistent with elevated expression of lysosomal-like enzymes and anti-microbial responses identified by Mydlarz et al. (2009). Ubiquitin ligase expression in this study may represent the nexus between the microbial diversity identified by Closek et al. (2014) and the anti-microbial responses identified by Mydlarz et al. (2009).

Expression profiles enhance findings of previous CYBD studies

Results from this study show that transition tissue 2–4 cm away from a visible lesion exhibits a different expression profile compared to visible lesion tissues (see Figs. 2–4). These results reaffirm conclusions of previous studies (Mydlarz et al., 2009; Weil, Cróquer & Urreiztieta, 2009b; Anderson, Armstrong & Weil, 2013; Closek et al., 2014) that characteristic differences exists between healthy and diseased colonies, as well as differences between diseased tissues and asymptomatic tissues that precede visible lesions on diseased colonies.

Results in this study coupled with Closek et al. (2014) findings that microbial diversity was highest in tissues the precede visible lesions raise the intriguing possibility that pathogen(s) responsible for CYBD may reside within the microbial community of the transition tissues. If the etiological agent for inducing CYBD actually resides in the microbial community of the transition tissues, then disease progression through a colony may actually follow a pattern similar to secondary succession observed in terrestrial communities after a disturbance such as a fire. The visible demarcation between transition tissues and lesion tissues corresponds to “fire line” of the disturbance. Consequently, lesion tissues behind the “fire line” are committed to subsequent degradation. Therefore the Symbiodinium response in transition tissues represents a potential sentinel of the approaching disturbance.

Mydlarz et al. (2009) investigated the anti-oxidant responses of prophenoloxidase (PPO) and peroxidase (POX) in CYBD tissues. Without the presence of significant ROS, down-regulation of PPO and POX in CYBD tissues would be expected when amounts of reactive oxygen species (ROS) generated is minimal. As previously identified, there was no significant up-regulated expression of SOD in coral host tissues (data not shown) which is congruent with the down-regulation of the antioxidant enzymes quantified by Mydlarz et al. (2009) and the absence of SOD expression identified by Closek et al. (2014). Concurrently, the mitochondrial response of elevated cytochrome oxidase further suggests a potential link between the absence of detectable ROS and the metabolic process of photorespiration which integrates the functions of mitochondria, chloroplasts, and peroxisomes through the photorespiratory C₂ cycle (Bauwe, Hagemann & Fernie, 2010; Voss et al., 2013).