Genome dynamics in Bacillus megaterium

1
Genome dynamics in Bacillus megaterium
What genomic sequencing tells us about the
genetic forces that shape Bacillus genomes

• October 29, 2009
• Dept. of Biological Sciences
• NIU

2
The Genus Bacillus

• Gram-positive, aerobic endospore-forming
  rod-shaped bacteria
• Normal habitat the soil (plus lots of other
  places)
• Mostly mesophilic, but some grow as low as 0and
  as high as 65.
• Pathogens B. anthracis and B. cereus
• Industrial uses enzyme production, Bt
  insecticidal corn
• Endospores very resistant to heat and chemicals

3
Relatives among the Firmicutes
4
A Bit of History

• Bacillus subtilis, originally named Vibrio
  subtilis, by Christian Gottfried Ehrenberg in
  1835. He was the first to use the name
  bacteria.
• Ferdinand Cohn (1872) renamed the species
  Bacillus subtilis, as part of his description of
  bacteria by their shape (bacillus little
  stick).
• --He is also responsible for bacteria being
  considered plants and not animals
• Robert Koch first showed that a specific
  bacterium caused a specific disease B. anthracis
  and anthrax. (1876)
• B. megaterium was first described by Heinrich
  Anton de Bary in 1884.

5
Bacilluss Position in the Tree of Life

• Anything called Bacillus in the 1800s would
  now be a member of the Firmicutes (strong
  skin), a phylum that contains the Gram-positive
  low GC bacteria.
• An alternative model, based on indels in
  universal genes, puts the Firmicutes near the
  root of the tree.

6
Bacillus Taxonomy

• Bacillus is a very old genus name, and it has
  been split several times.
• Bergeys Manual of Systematic Bacteriology, first
  edition (1986) lists 32 valid species, with about
  an equal number of synonyms.
• Based on morphology, biochemistry, some DNA-DNA
  hybridization, numerical taxonomy
• Carl Woese introduced the use of 16S rRNA
  sequences for phylogeny in 1977.
• Bergeys Manual second edition (2004) splits the
  Bacillus genus into 4 families, with 37 genera in
  the Bacillaceae. Over 200 species.
• Bacillus is still a genus, and still contains
  both B. subtilis and B. megaterium.
• As in other taxa, a common phenotype is well
  correlated with a common genotype

7
Ash et al. (1991) Lett. Appl. Microbiol.
13202-206.
8
Genome Sequencing

• Strain QM B1551, containing 7 plasmids
• NSF Grant, to Pat Vary and Jacques Ravel
• Most lab work done at TIGR/U. Maryland
• NIUs role annotating the 6000 genes
• Joined forces with Dieter Jahns group at the
  Technische Universität in Braunschweig, Germany,
  who were sequencing the DSM 319 plasmidless
  strain
• In addition, there are about 20 other fully
  sequenced genomes from Bacillus and related
  genera
• DSM319 has no plasmids, but at least 70 genes on
  the QM plasmids have good homologues on the DSM
  chromosome (purple ring near the middle)

9
Common Features, Genetic Forces
Assuming that all Bacillus species descended from
a common ancestor, what is similar and different
between them, and why?

• Genetic Forces
• Vertical descent
• Background substitution and indel mutations
• Horizontal gene transfer (about 10 different
  genes between QM and DSM)
• Intragenomic recombination
• Homogenization of rRNA operons, presumably by
  gene conversion

• Common Features
• Morphological and biochemical characteristics
• 16S rRNA genes
• A group of common protein-coding genes
• Chromosomal synteny
• rRNA operons

10
16S Variation, Phylogeny, and Species
Identification

• B. megaterium has 11 rRNA (rrn) operons on the
  chromosome in both sequenced strains, in the same
  genomic positions.
• QM also has an rrn operon on plasmid pBM400,
  which is not found in DSM.
• The 16S genes in B. megaterium are 1540 bp long
  and very similar, but not identical.
• Gene conversion is thought to homogenize rRNA
  operons
• Recombination between rrn operons leads to
  deletions
• The question addressed here what effect does 16S
  variation within the genome have on phylogeny and
  species identification?

11
Differences between 16S genes with B. megaterium

• Seven identical 16S genes the rrnE, rrnF, and
  rrnI genes in QM and the rrnA, rrnB, rrnF, and
  rrnK alleles in DSM.
• Also, the rrnA and rrnB alleles in QM were
  identical to each other
• Note the lack of clear vertical descent in this
  pattern
• Total of 20 sites with polymorphisms.
• All but 4 are unique to a single operon
• All but one shared polymorphism are found in both
  QM and DSM

• Positions 461 and 474 are probably a stem-loop
• all genes with an A at 461 have a T at 474, and
  all lines with a G at 461 have a C at 474.

12
Mismatch Differences in Completely Sequenced
Genomes
13
Differences in Completely Sequenced Genomes

• Maximum differences within any genome 16 (B.
  clausii)
• My basic argument there is no point in having
  two different species which are less different
  than 16S genes within the same genome.
• Among the cereus group genomes, there are fewer
  differences between genes in B. cereus, B.
  anthracis, and B. thurengiensis, than there are
  between genes in the same genome.
• Also, B. weihenstephanesis has only very few
  differences from these
• B. subtilis and B. amyloliquifaciens are also
  very similar.
• Effects on phylogeny pick a random 16S gene from
  each genome, align, count differences, do a
  neighbor-joining tree. 1000 reps.

14
Neighbor-Joining Trees with Completely Sequenced
Genomes

• Different choices of which 16S genes to use leads
  to different phylogenies, both at the
  species/subspecies level and at higher levels.
• The variable nodes in the cereus group and the
  halodurans/clausii group are independent. Thus,
  these three tree represent 9 variants.

15
Defining B. megaterium and distinguishing it from
other species, using 16S

• Comparison of B. megaterium isolates from Genbank
  to QM-rrnA
• A total of 185 isolates that were gt1390 bp (i.e.
  gt 90 of full length) and had fewer than 10
  ambiguity characters were aligned with QM_rrnA,
  and the number of variant positions were counted.
• 70 have 9 or fewer differences
• 86 have 20 or fewer differences
• 95 have 46 or fewer differences.
• Most isolates seem to fall into a single group,
  but there may be some significantly different
  subtypes in B. megaterium.
• Or, new species may be defined

16
Positions of Nucleotide Variants in Genbank
Isolates

• 43 of the 1540 nucleotide positions in the 16S
  gene have at least one variant in the B.
  megaterium strains from Genbank.
• Most of the variation occurs at the ends of the
  16S gene. This is also the region where missing
  data is most common.
• PCR primers for 16S need to be internal to the
  gene
• The variant positions in the middle were seen in
  QM and DSM the paired 461/474 positions, and
  1140. There are no major polymorphisms outside
  the end regions that are not seen in QM and DSM.

17
Closely Related Species

• How easy is it to distinguish between B.
  megaterium and closely related species?
• What species are closely related to B.
  megaterium? Different phylogenetic trees give
  different answers.
• All of the species on the next slides appear to
  be more similar to B. megaterium than members of
  the cereus group on at least one phylogenetic
  tree.
• All are in genus Bacillus except Lysinibacillus
  sphaericus.
• Total of 344 strains used

18
Number of Differences from QM-rrnA for Different
Species

• Except for B. flexus and one B. simplex isolate,
  all strains are well-differentiated from B.
  megaterium with a minimum of 58 differences.
• B. flexus overlaps the B. megaterium distribution
  heavily. The average B. flexus isolaate had 29.4
  differences from QM_rrnA, with some isolates
  indistinguishable. Type strain differs at 16
  positions the B. megaterium type strain differs
  at 4 positions.
• The average B. simplex isolate had 79.4
  differences from B. megaterium the one
  exceptional strain had 17 differences (maybe its
  a mis-labeling)

19
Conclusions about 16S genes

• Choosing different 16S genes from within genomes
  can affect the resulting phylogenetic trees.
• The 16S genes within B. megaterium and other
  completely sequenced Bacillus genomes differ from
  each other by up to 16 positions.
• Some species differ from other species at fewer
  positions than 16S genes differ within individual
  genomes.
• Although most B. megaterium strains are very
  similar to QM and DSM, there are a few strains
  with very different 16S genes that may represent
  subtypes within B. megaterium, or which may
  ultimately be assigned to new species.
• Most of the polymorphisms in the 16S genes are
  almost unique all of the widespread . megaterium
  polymorphisms are found in QM and DSM.
• Most of the closely related species fall outside
  the range of variation seen within B. megaterium,
  but B. flexus is a major exception.
• some isolates of B. flexus are indistinguishable
  from B. megaterium, and most fall within the same
  range of variation seen in B. megaterium

20
Common Genes and Synteny

• Bacillus is a relatively well-sequenced genus,
  with 11 complete genomes publicly available (not
  including B. megaterium).
• What genes are found in all Bacillus species, the
  core genome?
• Where on the chromosome are the conserved genes?

21
Bacillus Core Genome
22
QM vs. DSM Genes
23
Between Species
24
Synteny Results

• The syntenic region around the origin of
  replication is shared throughout the Bacillaceae,
  including the genera Bacillus, Geobacillus,
  Oceanobacillus, and Anoxybacillus.
• 99 of the 2000 core genes are in the syntenic
  region.
• Next rRNA operons and adjacent genes concrete
  examples of conserved synteny.

25
rRNA operons (rrn)

• There are 11 rRNA operons on the B. megaterium
  chromosomes, plus one on plasmid pBM400 in the QM
  strain.
• Other Bacillus species have 8-15 rrn operons
• The rrn operons are in the conserved synteny
  region.
• Only in Bacillus and relatives
• rrn operons are all on the leading DNA strand
  transcribed in the same direction as the
  replication fork moves.
• Most Bacillus rrn operons are on the right
  replichore, near the origin of replication

26
From Stewart and Cavanaugh, 2007, J. Mol.
Evol. 6544-67
27
Common Sites

• Nearly all the rrn operons in the Bacillaceae can
  be found between sets of common flanking genes.
• Sometimes with DNA insertions separating the rrn
  locus from one side
• A few unique rrn operons, including 2 in B.
  megaterium
• Not in Paenibacillus

A DNA repair protein recF B DNA gyrase, subunit
B C DNA gyrase, subunit A D
inosine-5'-monophosphate dehydrogenase E
D-alanyl-D-alanine carboxypeptidase. F glutamine
amidotransferase, synthase subunit
28
rrn operons in Bacillaceaae are in specific sites
29
Variations

• Seven sites on the right replichore, plus one on
  the left replichore.
• Also, a site shared within the cereus group, and
  two sites shared in Geobacillus and
  Anoxybacillus.
• Individual rrn sites can contain 0-5 rrn operons.
• Some sites are empty the flanking genes are
  adjacent, with no rrn operon between
• A few sites are missing the flanking genes are
  not present in the genome or are dispersed to
  very different locations.
• Tandem duplications of rrn operons are common
• Several variations caused by apparent
  intragenomic recombination

30
Tandem Duplication Copy Number
31
Intragenomic Recombination at rrn Sites

• rrn operons are almost identical, among the very
  few repeated sequences in bacterial genomes
• A second example insertion sequences (IS) ,
  which are mobile genetic elements found in many
  genomes (very few in B. megaterium ).
• The presence of highly conserved genes and the
  consequences of intragenomic recombination in a
  circular genome constrains genome rearrangements.
  
• The arrangement of rrn operons and their sites
  can be understood as the result of three forces
• intragenomic recombination between rrn operons,
• insertions/deletions of blocks of protein-coding
  genes,
• recombination events within tandem arrays of rrn
  operons.

32
Symmetrical Inversion Between Replichores

• Anoxybacillus flavithermus

33
Double Crossover Re-orders Flanking Genes

• B. pumilis

34
Double Crossover in Flanking tRNA Regions

• B. amyloliquifaciens rrnE.
• Resulted in loss of 2/3 of the 16S gene.
• 23S and 5S OK
• very little obvious homology on the right side.

35
Tandem Duplication Events Duplication by
Unequal Crossing Over

• rrnD in Oceanobacillus iheyensis

36
Tandem Duplications in the cereus group rrnG site

• Every deletion between adjacent rrn operons can
  be seen.
• Deletion of genes between rrn 2 and rrn3
  (preserving one gene in the middle).
• Region between rrn 3 and rrn 4 completely
  replaced.

37
Intragenomic Recombination Conclusions

• Most rrn operons are found in the same sites in
  all Bacillus genomes
• Differences in rrn operon number are mostly due
  to tandem duplications within these sites
• Intragenomic recombination is well documented in
  Bacillus genomes
• Anoxybacillus symmetric inversion across ori
• B. pumilis double crossover involving 3 regions
• Oceanobacillus rrnD CO between tandem copies
• rrnD in other species at least 2 events
• cereus group rrnG deletions between tandem
  copies (at least 4 different events)
• cereus group rrnG replacement of inter-rrn
  region by presumed 2CO
• cereus group rrnG deletion of inter-rrn region,
  leaving a central portion intact (2 deletions?).
• B. amyloliquifaciens rrnE 2CO involving 3
  regions, with the central section having the COs
  150 bp apart
• B. megaterium rrnBC 2CO involving 3 regions,
  with little homology at one end
• Several other duplication/deletion events within
  tandem duplications

38
Some Events NOT Observed

• The lack of certain events supports several
  current ideas.
• to the extent that lack of evidence constitutes
  evidence.
• Crossovers between rrn sites despite numerous CO
  events within rrnG in the cereus group, plus many
  other CO events
• supports the idea that the flanking genes are
  necessary
• Asymmetric CO across ori only one symmetric one
  observed, so evidence is not strong.
• Supports the idea that symmetric replichores are
  selectively advantageous
• Inversions within a replichore all rrn in all
  species are on the leading strand, in both
  replichores.
• supports the idea that replication and rrn
  transcription must proceed in the same direction

39
Some Unsolved Questions

• Replichore asymmetry
• most of the rrn are in the right replichore
• compositional bias between replichores
• Mechanism of insertion/deletion/horizontal gene
  transfer
• a big question. We are examining insertion sites
  for clues.
• Is there a common phylogeny for the conserved
  synteny region in B. megaterium?
• Finding and analyzing allegedly unique events
  (indels and recombinations)

40
DNA Composition shows replichore asymmetry
41
Simple vs. Compound Insertions
42
Thanks!

• NIU Biology Dept.
• Pat Vary
• Janaka Edirisinghe
• Kirthi Kumar Kutumbaka
• Sandhya Balasubramanian
• Jenn Hintzsche
• Chris Braun
• Denise Tombolato
• Judy Luke
• Scott Grayburn
• NIU Computer Science Dept.
• Stephen Snow
• Reva Freedman
• Minmei Hou

• Argonne National Labs
• Ross Overbeek
• Gordon Pusch
• Terry Disz
• TIGR/U. Maryland
• Jacques Ravel
• Mark Eppinger
• MJ Rosovitz
• Technische U. Braunschweig
• Dieter Jahn
• Boyke Bunk