Materials and Reagents

The main starting materials for embarking upon a large TFome project are the assembly of gene models and a collection of plasmid templates from existing cDNA collections. In this section, processes to develop these are outlined.

1. Assembly of gene models and TF domain databases
   
   1. Assemble a collection of gene models from the available target genome. For plant genomes, the Phytozome database (phytozome.jgi.doe.gov) and EnsemblPlants (http://plants.ensembl.org/index.html) permit the downloading of all predicted protein models for a species as a single multi-sequence FASTA formatted file. For example, at the Phytozome website the BioMart tool permits one to select the genome dataset for a particular plant species. Once the genome is selected then attributes of the sequences to be downloaded can be specified such as peptide sequences or coding sequences only. By default all gene models are selected and then these can be downloaded as a single FASTA file that is the input file for subsequent steps below.
      
   2. Assuming the draft quality of the transcript annotation it would be desirable to eliminate redundant gene models in a multiseq FASTA file. The GRASSIUS website provides the custom perl script “IdentifySeqRedundancy.pl” (http://grassius.org/tools.html) for this purpose. This script also requires the perl module “Digest::MD5” which is available from the Comprehensive Perl Archive Network (CPAN) website (http://search.cpan.org/dist/Digest-MD5/MD5.pm). One can also eliminate redundant models within the species using BLAST searches. Proteins were arbitrarily considered duplicated if they are found in the same species, with a query coverage ≥ 90%, or have a query identity ≥ 90% and the query alignment starts less than 9 residues from the start codon. Alternatively, more complex criteria such as those of Gu et al., 2002, may be employed to identify duplicate proteins in a genome. If these conditions are satisfied, the longest protein is kept and the eliminated proteins were classified as identical or splicing variants. If there is access to RNA-Seq datasets they may be used to corroborate target TF gene models, but such an approach is not outlined here. Targeting the longest splice variant which is supported by EST or RNA-Seq data provides the maximum protein interaction space for identifying the protein-DNA and protein-protein interactions that gene regulatory networks are comprised of.
      
   3. Scanning the non-redundant multifasta file against a collection of protein domains as hidden Markov models (HMMs) such as provided by PFAM or Interpro provides a protein domain annotation of the putative proteome. For this particular step the software HMMER is required. In brief HMMER is a set of tools implementing the profile hidden Markov models to find similarity across protein sequences and is necessary to search the PFAM HMM models in a group of protein sequences. It is possible to call HMMER from a variety of scripting languages such as perl or python, using pre-build HMMER wrappers. PFAM provides one of those scripts written in perl (pfamscan.pl) with a set of PFAM pre-established parameters. Source code may be downloaded from the following website ftp://ftp.sanger.ac.uk/pub/databases/Pfam/Tools/PfamScan.tar.gz. Table 1 is a compilation of protein domains that have been used to define TF and CR families in plants (Buitrago-Florez et al., 2014; Burdo et al., 2014; Perez-Rodriguez et al., 2010; Yilmaz et al., 2009). This table describes 62 TF and 26 CR families that were used in defining gene models for inclusion in the maize TFome. Twelve of these families do not have ascribed PFAM domains but can be defined using “in house” HMMs as previously described (Buitrago-Florez et al., 2014; Perez-Rodriguez et al., 2010). The HMMs for the CCAAT-Hap2, 3, and 5 TF families can be obtained from the AGRIS database (http://arabidopsis.med.ohio-state.edu) (Yilmaz et al., 2011).
      
   4. Once the primary list of gene models are annotated with protein domains, they are assigned to TF or CR families using the criteria outlined in Table 1. The criteria include identifying the presence of one or more motifs and the absence of other (forbidden) motifs (Buitrago-Florez et al., 2014; Perez-Rodriguez et al., 2010; Yilmaz et al., 2009). The annotation involves a custom perl script “get_InterProScanDomains.pl” to sort proteins into TF families based on the interproscan output and is available at the GRASSIUS website (http://grassius.org/tools.html). The script keeps only significant hits with e-value ≤ 0.001 and verifies that the rules as defined in Table 1 are fulfilled. When the rules are only partially fulfilled then the protein is assigned into “Orphan” TFs. This can be the largest class of TFs for a species (~11.7% of the maize TFome), until Orphan members are assigned to families.
      
   5. Searches can be conducted with local computation resources or when not available using the iPlant Collaborative Discovery Environment public resource (http://www.iplantcollaborative.org).
   
   
2. Screening of EST or flcDNA collections for plasmid templates
   
   1. Identify which cDNA resources are available for the target genome (see Note 1). Most genome projects for model species include generating an EST library, although many libraries are not publically available. For the maize TFome, extensive use of the maize flcDNA collection (http://www.maizecdna.org) was made (Soderlund et al., 2009). Individual cDNA clones for this library and many other plant species are available through the Arizona Genomics Institute (http://www.genome.arizona.edu).
      
   2. Using the coding sequence of a gene model as a query sequence, ESTs or flcDNAs were then identified using BLASTP or BLASTX, for which the amino terminus, including the start codon, was present and the sequence alignment was >99%. Alignments that are not 100% identical may occur because EST sequences are not high fidelity due to sequencing errors particularly at the 3’ end of sequences (Soderlund et al., 2009). Alternatively spliced isoforms not represented by a transcript model may also be targeted, as long as they maintain the reading frame of the original transcript.
      
   3. Once a likely flcDNA template is located, then the plasmid is acquired, isolated and sequenced from each end to confirm that: a) it is the correct template, and b) it is full length. Some cDNA libraries utilize enzymes during their construction that cut within the coding sequence leading to partial clones, which would be unsuitable for amplification of the entire coding sequence.
      
   4. Once a suitable target plasmid template is identified for a target gene, then a sample is diluted to 1 ng/µl and 1 ng is sufficient as template in a PCR reaction.
   
   
3. Other reagents
   
   1. One Shot^® TOP10 chemically competent cells (Life Technologies, catalog number: C404003 )
      
   2. PureLink^® Plant RNA Reagent (Life Technologies, catalog number: 2322-012 )
      
   3. Turbo DNA-free™ kit (Life Technologies, Ambion^®, catalog number: AM1907 )
      
   4. Thermo Scientific Maxima H minus 1^st strand synthesis kit (Thermo Fisher Scientific, catalog number: K1652 )
      
   5. Ribolock^® RNase inhibitor (Thermo Fisher Scientific, catalog number: EO0382 )
      
   6. EmeraldAmp^® MAX PCR Master Mix (Takara Bio Company, catalog number: RR320A )
      
   7. Phusion^® high fidelity polymerase (New England Biolabs, catalog number: M0530S )
      
   8. Gateway^® pENTR™/D-TOPO^® or pENTR™/SD/D-TOPO^® vectors (Life Technologies, catalog numbers: K243520 and K242020 respectively)
      
   9. Genscript^® Taq Polymerase (GenScript USA Inc., catalog number: E000071000 )
      
   10. EmeraldAmp^® MAX PCR Master Mix (Takara Bio Company, catalog number: RR320A)
       
   11. 1 kb ladder molecular weight size standards (Thermo Fisher Scientific, catalog number: SM0311 )
       
   12. RNA extraction buffer I (see Recipes)
       
   13. RNA extraction buffer II stock solutions (see Recipes)
       
   14. RNA extraction buffer II (working solution) (see Recipes)
       
   15. Carlson lysis buffer (see Recipes)
       
   16. Freezing medium (see Recipes)

Equipment

Note: Most equipment listed here is standard molecular biology instrumentation present in most laboratories. A large TFome project would also benefit from the use of multichannel pippetters and 96 well plate formatted experimentation.

1. Microcentrifuge
   
2. High speed centrifuge
   
3. Gradient thermocycler
   
4. Gel electrophoresis equipment
   
5. Water baths
   
6. -80 °C freezer for the storage of stocks
   
7. Aerosol pipette tips recommended for all DNA manipulation and cloning experiments
   
8. Wizard^® SV 96 Plasmid DNA Purification System (Promega Corporation, catalog number: A2255 )
   
9. Wizard^® Plus SV Minipreps DNA Purification System (Promega Corporation, catalog number: A1330 )
   
10. Wizard^® SV Gel and PCR Clean-Up System (Promega Corporation, catalog number: A9281 )
    
11. Gene synthesis (Life Technologies)
    
12. 2 ml 96 well culture dish (Thermo Fisher Scientific, catalog number: 12566121 )
    
13. Breathable plate seal (Thermo Fisher Scientific, catalog number: AB0718 )
    
14. 0.5 ml plates (USA Scientific, catalog number: 18965000 )
    
15. 7 mm sized silicone seal that can be re-autoclaved (Thermo Fisher Scientific, catalog number: 0339649 )
    
16. Matrix sample storage system (matrixtechcorp.com, Thermo Fisher Scientific, catalog numbers: 4111MAT ) (Capit-All^® Capper/Decapper), 3740 (Barcoded screw cap tubes), and 4477 (Screw cap tray)
    
17. (Optional) Vac-Man^® 96 Vacuum Manifold (Promega Corporation, catalog number: A2291 ) for 96 well plasmid preparation
    
18. (Optional) Vacuum pump capable of generating 38-51 cm of Hg or equivalent
    

Software

1. OligoAnalyzer 3.1 software (www.idtdna.com)

Procedure