2.3.1 Plant material for reference transcriptome
Bulbs of Narcissus pseudonarcissus L. var. Carlton were obtained from Alzeim
Ltd. The supplier to Alzeim was New Generation Daffodils
(http://www.newgenerationdaffodils.com). Thirty bulbs of Carlton were planted in pots (36 cm diameter, 27.5 cm depth), 15 bulbs per pot on 14/12/2010. The bulbs were grown in a mixture of John Innes number 3 soil (Keith Singleton Horticulture, Cumbria, UK) and 3.0-6.0 mm Perlite supercoarse (William Sinclair horticulture Ltd, Lincoln.). Bulbs were planted at a depth of 15.2 cm. The pots were placed on the Institute roof thus experiencing normal weather conditions and inspected regularly, watering as necessary.
2.3.2 Basal plate extraction
Whole plants were dug up in mid April 2011 after the foliage had died back as this is the time predicted to show the highest level of galanthamine in the bulbs (personal communication with Dr X Chang). Four plants were washed in cold water and any remaining foliage removed. The roots were then removed and the basal plate cut from the bulb. This was then cut into small pieces of 1-2 mm thick, frozen in liquid nitrogen and stored at -80 °C until required. The basal plate was used in order to avoid the high levels of chloroplast transcripts expected in the bulb tissue causing rare or lowly expressed transcripts to be missed in the resulting sequencing.
2.3.3 Production of cDNA library
2.3.3.1 Total RNA extraction – CTAB method
The method was modified from Chang et al, (1993). The frozen basal plate tissue was ground in a pestle and mortar under liquid nitrogen. Extraction buffer (2% CTAB, 2% PVP40, 100mM Tris-HCl, 25mM EDTA, 2M NaCl, 0.5g l-1, 2% β-mercaptoethanol) was warmed to 65°C and 2g powdered frozen tissue added before vortexing until homogeneous. The solutions were then incubated for 20 min at 65°C then placed on ice. The solutions were extracted 3 times in 10ml chloroform:isoamylalcohol (24:1). The phases were separated via
centrifugation at 3696 g (Sorvall Legend RT with round buckets 75006533) at 4°C for 15 min carrying the supernatant (top layer) through between each extraction. Then 2.5 ml 10M lithium chloride was added to the supernatant and vortexed. The nucleic acid precipitate was allowed to form overnight at 4°C followed by centrifugation at 3696 g (Sorvall Legend RT with round buckets 75006533) for 30 min at 4°C. At this stage the RNA was purified using the RNeasy Plant mini kit (Qiagen) following the manufacturers RNA clean up protocol and dissolved in RNAse free water.
To determine the quality and quantity of RNA, 1µl was analyzed spectrophotometrically (Labtech NanoDrop® ND-1000 spectrophotometer). This determined concentration and OD ratios 260:280 and 260:230. The RNA was then analysed for degradation using formaldehyde gels following the method of Chang et al., (1993).
2.3.3.2 mRNA isolation
Two different protocols were used to isolate mRNA from the total RNA preparation to determine whether mRNA depletion or rRNA selection produced the higher yield and quality of mRNA for cDNA library preparation.
2.3.3.3 rRNA depletion
Sample 1 (concentration 10 µg total RNA in 1.1 µl) was rRNA depleted using the Ribominus kit (Invitrogen) according to the manufacturer’s protocol. The method works by hybridizing rRNA molecules to locked nucleic acid probes for known rRNA molecules 138.
2.3.3.4 mRNA selection
Sample 2 (45 µg total RNA in 5 µl) was mRNA selected using the Dynabeads mRNA purification kit (Invitrogen) according to manufacturers protocol. This method involves the pairing of the poly A chains on the 3’ end of mRNA to the oligo (dT)25 residues on the surface of the beads 137.
2.3.3.5 cDNA Library preparation
Following rRNA depletion (sample 1) or mRNA selection (sample 2), both were
used to construct cDNA libraries following the manufacturer’s
recommendations for the GS FLX titanium series cDNA Rapid Library Preparation Method (Roche). Both preparations started with 200 ng RNA. Since the samples would be pooled for sequencing, adaptor ligation was carried out to barcode the separate samples. Sample 1 was ligated to RL MID6, ATATCGCGAG, (Roche) and sample two to RL MID7, CGTGTCTCTA, (Roche).
The libraries were analyzed to check that the rRNA had been removed and that the fragmentation had been successful after depletion, fragmentation and at the end of the preparation using Agilent RNA 6000 Nano Kits. The resulting libraries were stored at -80°C until sequencing.
2.3.4 Creation of a reference transcriptome
2.3.4.1 454 Pyrosequencing
Both libraries were sequenced at the Centre for Genomic Research (CGR) at the University of Liverpool. This was carried out on the Roche 454 Pyrosequencing Titanium FLX series instrument. The libraries were amplified independently then pooled on half a plate for sequencing.
2.3.4.2 Assembly of reads from sequencing of barcoded libraries
The resulting sff files from the 454-pyrosequencing runs were assembled using the GS De Novo Assembler Newbler program (version 2.5). The program was run using the default settings. The two sff files corresponding to the barcoded libraries were first assembled independently and then assembled together.
2.3.4.3 Re-sequencing of sample 2 and full assembly from all GS FLX data
Sample 2 was re-sequenced at the CGR to give a greater coverage of the transcriptome due to the low read number of the original sequencing run.
A further assembly was carried out using the GS De Novo Assembler Newbler program (version 2.5). This assembly will be known as the full assembly from here on.
2.3.4.4 Manual annotation and the creation of an automated annotation pipeline for the joint and full assemblies using “full_annotation.pl”
The two assemblies were annotated separately using the same methodology as described below. The annotation was implemented using a variety of custom- built perl scripts and command line. The steps involved are shown in figure 2.1. The individual steps in the pipeline were combined to produce a single script that ran the whole pipeline from the command line in one step.
Figure 2-1 Annotation pipeline using "full_annotation.pl".
The steps were implemented using perl scripts and command line prompts. The eventual aim would be to have this as a simple pipeline that can be used by non-bioinformaticians for future analysis. The script “full_annotation.pl” can be seen on appendix disc. The steps in blue were carried out using the command line version of BLAST 2.2.27+, using BLASTX for steps 2,4 and 6 and BLASTN for step 3 with the –m 8 tabulated output option and –b 1 and –v 1 options for upper limits on number of database sequences to show alignments for and number of one-line descriptions to show. The steps in red were carried out via a perl script created by Richard Gregory of the CGR at the University of Liverpool that removes high scoring pair results so that each transcript has only one hit. The steps in green were implemented using a variety of UNIX commands and perl hashes to pull out the sequences from the fasta file that had no hits.