PCR for species with extreme GC-content

In spite of recent development of deep sequencing technologies, we still sometimes have to isolate molecular sequences in targeted searches especially for non-model organisms.

I am writing this to share info which I and my colleagues have collected to tackle difficult RT-PCR targets in everyday research projects.

In fact, most of my struggle was with lampreys (Petromyzon, Lethenteron, Lampetra), some of which you can see in high GC-content, unique codon usage pattern and amino acid composition of KCNA genes recently published by my team (Qiu et al., 2011).

1. PCR for species with high GC-content

Example: lampreys, some choanoflagellates, amphioxus, amniote GC-rich isochore regions 

Typical problem: mRNA secondary structure in reverse-transcription preventing extension of first strand synthesis

Items to consider: 
A. Thermostabilization of reverse-transcriptase by tweaking reaction condition with trehalose (Carninci et al., 1998)
B. Use of thermostable reverse-transcriptase (e.g., ThermoScript, Invitrogen)
C. Modification of entire reverse-transcription protocol (Shi & Jarvis, 2006) using a special reverse-transcriptase Thermo-X (Invitrogen) (But, in Germany at least, Thermo-X is not available anymore)
D. Using a DNA polymerase and buffer optimized for GC-rich templates in PCR (GC-rich PCR system, Roche)

2. PCR for species with low GC-content

Example: some teleost fishes, planarians, (some?) cnidarians, amniote GC-poor isochore regions 

Typical problem: mispriming by oligo dT primers at A-rich region (but not at genuine poly-A tail) resulting in truncated 3'UTR (I call this 'pseudo-polyA')

Further reference:
How to distinguish polyadenylation signals (mostly AAUAAA or AUUAAA?) (Lee et al., 2007)

Items to consider:
A. Optimization of priming condition using trehalose (Mizuno et al., 1999)
B. When you see multiple bands after 3'RACE PCR, sequence as long fragments as possible