5.2. Comparative genome analysis within the species Paenibacillus larvae using suppression subtractive hybridization

Suppression subtractive hybridization (SSH) is a powerful tool for elucidating genomic sequence differences among closely related bacteria. SSH is a PCR-based DNA subtraction method which was originally developed for generating differentially regulated or tissue-specific cDNA probes and libraries (Diatchenko et al., 1996). However, it has also been successfully adapted for bacteria, especially for the identification of genes that contribute to the virulence of bacterial organisms (Akopyants et al., 1998). For example, SSH has been successfully employed to compare genomes of pathogenic and non-pathogenic (Janke et al., 2001; Reckseidler et al., 2001) or virulent and avirulent strains of bacterial pathogens (Zhang et al., 2000). SSH has also led to the identification of pathogenicity islands (Hacker et al., 1997) in infectious bacteria (Agron et al., 2002). Recently, SSH analysis of all four genotypes of P. larvae led to the identification of putative virulence factors like potent antibiotics belonging to the class of non-ribosomal peptides and polyketides as well as several toxins and cytolysins (Fünfhaus et al., 2009).

The principle of any SSH analysis is that one genome putatively containing additional genes (the so-called ‘driver’) is subtracted from the other genome (‘tester’) with the result that  the additional sequences specific to the tester remain and can be visualized via PCR. For SSH analysis, genomic DNA of P. larvae grown in liquid culture of brain heart infusion broth (BHI broth, see section 3.1. for recipe) is used. Since the quality of the genomic DNA is of crucial importance, special kits for DNA extraction of Gram-positive bacteria (e.g. MasterPure Gram Positive DNA Purification Kit, Epicentre Biotechnologies) to isolate P. larvae DNA are recommended. The following protocol for SSH is based on the protocols published by Akopyants et al. (1998) using the PCR-Select Bacterial Genome Subtraction Kit (BD Clontech). SSH consists of two consecutive phases, the hybridization and the amplification. Within the phase of hybridization, the genomic DNA extracted from the ‘driver’ strain is hybridized with DNA extracted from the ‘tester’ strain. Sequences that are present in the tester strain but missing in the driver strain are then isolated in the amplification phase in which target genomic DNA fragments are amplified, while  amplification of non-target DNA is simultaneously suppressed using the suppression PCR effect (Siebert et al., 1995).

         The protocol for the comparison of P. larvae genotypes involves the following steps:

  1. Following the manufacturer’s protocol, digest the isolated DNA from both genomes using restriction enzyme RsaI, which has a recognition sequence of only four bases.
    Since this sequence occurs often in the genome of P. larvae, a high fragmentation of the bacterial DNA (100 to 1000 bp) with blunt ends is achieved.
  2. After purification of the fragmented DNA with the MinElute Cleanup Kit (QIAGEN) according to the manufacturer’s protocol, divide the DNA of the tester strain into two pools which are ligated with either Adaptor 1 (5’–CTAATACGACTCACTATAGGGCTCGAGCGGCCGCCCGGGCAGGT-3’) or Adaptor 2R (5’–CTAATACGACTCACTATAGGGCAGCGTGGTCGCGGCCGAGGT-3’), catalysed by T4 DNA ligase.
    The ends of the adaptors lack a phosphate group, so only one strand of each can be ligated to the 5’ end of the tester DNAs.
  3. For the first round of hybridization, mix each pool of adaptor-ligated tester-fragments with a 50-100-fold excess of driver-fragments.
  4. Incubate the mixed samples at 98°C for 90 sec (denaturation).
  5. Incubate at 65°C for 90 min (annealing).
  6. In the second round of hybridization, mix both pools without denaturation.
  7. Incubate overnight at 65°C to allow free tester-fragments from both pools to form heterohybrids (hybridization of complementary tester DNAs with different adaptors (Akopyants et al., 1998)).
  8. Add freshly denatured driver to the mixture.
  9. Allow the samples to hybridize.
    During this step, hybrid molecules are formed, but only DNA fragments which are exclusively present in the sample of the tester strain are amplified in the subsequent amplification phase.
  10. For amplification, perform a nested PCR with adaptor specific primers.
    For the first PCR, the primer 1 (5'-CTAATACGACTCACTATAGGGC-3') is used. For the second PCR, the nested primer 1 (5'-TCGAGCGGCCGCCCGGGCAGGT-3') and the nested primer 2R (5'-AGCGTGGTCGCGGCCGAGGT-3') are used.
  11. To obtain a library of tester-specific sequences, secondary PCR products - i.e., clone tester specific DNA fragments - are then cloned into appropriate vectors (e.g. pCR2.1TOPO vector, Invitrogen, containing an ampR gene).
  12. Transform into competent E. coli (e.g. TOP10 cells, Invitrogen) following the manufacturer’s instructions.
  13. Plate transformation mixes on agar plates supplemented with Ampicillin (100 µg/ml)
  14. Incubate overnight at 37°C.
  15. Pick and grow clones individually overnight in Luria broth in the presence of Ampicillin (100 µg ml/ml) at 37°C and 300 rpm.
  16. Extract plasmid DNA from these overnight cultures (e.g. by using the Qiagen plasmid mini kit).
  17. The presence of inserted P. larvae DNA can be verified by restriction digestion. In the case of pCR2.1TOPO, perform digestion with EcoRI.
  18. Sequence detected inserts from positive clones using appropriate primers.
    For pCR2.1TOPO, use primers M13 uni (-21) 5’-TGTAAAACGACGGCCAGT-3’ and M13 rev (-29) CAGGAAACAGCTATGACC.
  19. To verify the specificity of the DNA fragments for the tester strain, perform a PCR with fragment specific PCR primers on DNA from the tester strain and related P. larvae strains (i.e., representatives of the same genotype).

         In each subtraction, all controls recommended by the manufacturer must be performed and all must test positive. Analysis of the tester-specific sequences is performed with BLASTx (Altschul et al., 1990, 1997) followed by functional annotation based on the COG (cluster of orthologous groups) classification (Tatusov et al., 1997, 2003).