Download Prokaryotic Genomics

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Gene expression wikipedia , lookup

Gene desert wikipedia , lookup

Lac operon wikipedia , lookup

Non-coding DNA wikipedia , lookup

Transcriptional regulation wikipedia , lookup

Community fingerprinting wikipedia , lookup

RNA-Seq wikipedia , lookup

Promoter (genetics) wikipedia , lookup

Genomic imprinting wikipedia , lookup

Gene regulatory network wikipedia , lookup

Silencer (genetics) wikipedia , lookup

Ridge (biology) wikipedia , lookup

Molecular evolution wikipedia , lookup

Gene wikipedia , lookup

Artificial gene synthesis wikipedia , lookup

Gene expression profiling wikipedia , lookup

Genome evolution wikipedia , lookup

Transcript
Genomics of bacteria and archaea:
the emerging dynamic view of the
prokaryotic world
E. V. Koonin and Y. I. Wolf
Nucleic Acids Research 36:6688-6719
October 2008
Extent of Prokaryotic Diversity
• Only about 0.1% of bacteria can be cultured in the
laboratory!
• Currently about 1200 sequenced prokaryotic genomes
• Large scale metagenomic surveys have not revealed
abundant bacteria outside of already known phyla
– Metagenomics = sequencing DNA found in the environment
without growing or purifying the organisms.
– Biggest survey: Craig Venter seawater survey.
– Only about 10% of metagenomic sequences have no discernable
homologs.
– Possibly many new species exist in some unusual habitats?
Genome Size
• Current smallest: Carsonella rudii = 180 kbp
• Current largest: Sorangium cellulosum = 13 Mbp
• Genomes less than 1 Mbp are all parasites or intracellular
parasites, which don’t need to make all compounds from
scratch.
– 1 Mbp seems to be about the minimum size for a fee-living bacterium
• The largest viruses (mimiviruses) are 1 Mbp or so; such viruses
are common in marine habitats.
• The smallest eukaryotic genomes (the obligate intracellular
parasite Encephalitozoon intestinalis) is 2.3 Mbp
Gene Density
• Roughly 1 gene per 1000 bp in
both bacteria and archaea
• Intergenic spaces are either almost
0 bp (within operons) or average
about 100 bp.
• Longer intergenic spaces probably
contain RNA-only genes or
pseudogenes
• Nearly all prokaryotic genes are a
single open reading frame, with
very few introns or split genes.
• Gene overlaps are no more than a
few base pairs: no documented
cases of long overlaps.
Clusters of Orthologous Groups of
Genes
• “orthologs” are genes that descend from the same gene in
an ancestral species.
– Need to be a bit looser in prokaryotes, where horizontal gene
transfer is common
– Often defined by “bidirectional best hits” (BBH): two genes (in
different genomes) are each other’s best blast hit in those
genomes.
– Problem of gene duplication: paralogues. Paralogues are also
derived from the common ancestor but have evolved different
functions.
• COGs are based on identifying orthologous genes, even if
there is more than one in a given genome.
• New derivative of COGs: EggNOGs (yuck). The database
includes genes from 312 bacterial species and 26 archaea.
COG results
• How widespread are different
orthologous gene families?
• In most sequenced genomes, about 80%
of genes can be assigned to a COG.
– The rest of the genes have no detectable
homology with any other protein; they
are often called “ORFans”
• There are very few COGs found in most
or all organisms (“core” genes: about 70
gene clusters)
• A larger, but still small number of COGs
is moderately conserved, found in many
genomes (“shell” genes: 5700 gene
clusters)
• The large majority of COGs are found in
only a few genomes (“cloud” genes:
24,000 clusters)
Out of the 338 genomes in EggNoG
how many are missing in each COG?
Same data in both plots, but the
bottom one is semi-log.
Percentage of Genes in EgGnOG
COGs
ORFans vs. ELFs
• A gene with no detectable homology to any other protein in another
species is an ORFan
• What are ORFans?
– Some are ELFs = Evil Little Fellows: falsely predicted genes; hypothetical genes
are aren’t real. (BTW--I don’t think ELF is going to make it into standard
genomics jargon, but ORFan might).
– Some are real genes derived from bacteriophages.
• Metagenomic studies suggest that the world bacteriophage genomes is vast and
very under-explored. This is a very important concept that we will explore later.
• In genome annotation, it is common to find prophages, which look like regions of
the genome with many hypothetical genes mixed with a few genes labeled “phage
protein” or “integrase/recombinase”
– Some are just the tail end of the distribution of the “cloud” genes that are
found in only a few genomes (in this case, just 1 genome).
• How big is “gene space”, the totality of all genes? Could be several orders of
magnitude larger than we know now, and almost all of it will be genes found in only
1 or a few genomes, or perhaps only in bacteriophage.
COGs in Phylogenetic Groups
• The presence or absence of members of each of the 30,000+
COG groups in all the 338 EggNog genomes can be used for
cluster analysis. (Self-organizing map, here).
– On the SOM, genomes close to each other share more COGs than
genomes far apart on the map.
– There is quite a correlation between COG presence/absence and
known phylogenetic groups (based on 16S rRNA): different members
of the same phylum group together.
– with a few exceptions: gamma-proteobacteria are split, possibly due
to a diversity of life styles.
COGs vs. Gene Function
•
Genome annotation is based on the principle that if someone experimentally
determines a gene’s function, then all other genes with similar protein sequences
perform the same function.
– Annotation also uses information about the gene’s chromosomal neighborhood: genes
that are part of the same subsystem are often found grouped together.
– We are not likely to be able to predict a protein’s function directly from its amino acid
sequence anytime soon.
•
Non-orthologous gene displacement is common. When two organisms are
compared, the same gene function is performed by two entirely different, nonhomologous proteins.
– This happens even in very fundamental processes like DNA replication: the primary
enzymes for replication are entirely different between the bacteria and the archaea.
•
Because of non-homologous gene displacement, the “gene sequence homology
space” in the previous slide is not identical to a SOM map of “gene function space”
– SOM set up the same way: vector of presence/absence of different gene functions
(functional roles) in different species
– Here, the phyla are less well grouped. Perhaps because even closely related bacteria
often have very different lifestyles.
Genome Architecture
•
•
Most prokaryotes have a single DNA origin of replication (ori), which is used to define base 1
in a genomic sequence, as well as the orientation of the sequence.
DNA polymerase starts replication at ori and goes in both directions, which defines a “leading
strand” (the right half of the genome) and a “lagging strand” (left half).
–
•
•
These can also be called the right and left replichores.
The two halves often have noticeably different base compositions (GC content, etc.).
Most genes, especially highly transcribed genes, are oriented in the same direction as
replication.
Dotplots to Compare Genome Structure
• Compare positions of orthologous genes between 2 genomes,
then plot positions.
– A. Closely related genomes are mostly collinear, or syntenic (this is two
Geobacillus species)
• syntenic means that neighboring genes in one species are also neighbors
in another species
– B. Moderately related bacteria show an X-shaped pattern due to
multiple inversions across the origin (which preserves the direction of
transcription). Shewanella
– C. X-pattern in 2 Archaea: Pyrococcus
– D. Distantly related species show a random distribution of orthologs:
genome is well-scrambled.
• In general, only closely related species show any common
genome architecture. The overall arrangement of genes on
the chromosome is not well preserved.
Bacillus Dotplot: B. megaterium vs. B. cereus
--organization is conserved in the vicinity of the replication origin, but
not in other regions.
Operons
• The classic operon is the E. coli lac operon.
– Jacob and Monod, 1962
– Three genes involved in lactose utilization are transcribed onto a single
messenger RNA
– Transcription is under the control of a single transcription factor, the
lac repressor.
– When the lac repressor detects lactose, it allows the operon to be
transcribed.
•
Most prokaryotes have numerous operons of many types
Operons Across Species
•
•
Operon structure is conserved much better than overall chromosomal synteny
especially for genes whose proteins physically interact, such as the ribosomal
proteins.
– Interpretable as selection for having a balanced number of all subunits.
•
The 50+ ribosomal proteins are found grouped in different patterns across all
prokaryotes. The ribosomal “superoperon”
– other groups of partially conserved operons also exist, giving the general concept of the
conserved gene neighborhood: even when they are not part of the same operon, genes
involved in the same subsystem tend to stay near each other.
•
•
However, most operons are not part of superoperons, but rather just 2-4 genes
that are oriented in the same direction and are co-transcribed and co-regulated.
Conservation is moderate: operon membership tends to change over phylogenetic
distance.
–
•
However, most groups of adjacent genes in the same orientation are actually co-regulated as
operons.
The percentage of genes in operons varies: very high in Thermotoga, very low in
Cyanobacteria.
Gene Regulation and Signal Transduction
• Lac operon model: a single protein senses something in the environment
(lactose) and directly alters transcription.
– Some variants: genes transcribed from a common regulatory region in
opposite directions (a divergon), and genes in multiple locations affected by
the same regulatory protein (regulon)
• The transcription factors (DNA binding proteins that affect transcription)
are well conserved, but which genes are affected varies widely.
– Transcription factors generally consist of a ligand-binding domain (e.g the part
that binds to lactose) and a DNA binding domain.
• Two component histidine kinase systems:
– one protein in a membrane-bound histidine kinase that senses something in
the extracellular environment.
– The histidine kinase phosphorylates another protein, the response regulator,
which is soluble and binds to the DNA to affect transcription.
• Many other systems, often originally found in eukaryotes: cyclic AMP,
cyclic di-GMP, programmed cell death systems, and more.
Genome Size
• Minimal number of genes:
– for growth on rich medium, where there are very few biosynthetic
requirements: maybe 250 genes.
• Carsonella rudii, an obligate intracellular parasite, has only 170 genes. It even lacks
some aminoacyl tRNA synthetases, and probably uses host enzymes for this
function. Perhaps it is being converted into an organelle? (that’s just speculation,
however)
– for a free living heterotroph, maybe 1000 genes are needed
• Pelegibacter ubique has about 1100 genes
– given the presence of non-homologous gene replacement and different
lifestyles, there are undoubtedly many more-or-less minimal genomes that
survive.
Gene Class vs. Genome Size
•
•
•
•
•
Some genes are found in about the same
numbers in all genomes: translation machinery,
cell division machinery.
Other genes are proportional to genome size:
metabolic genes, transporters, DNA replication
and repair
Other genes are proportional to the square of
genome size: regulatory proteins.
– Small genomes have very few regulatory
proteins, while large genomes have lots.
The fraction of regulatory genes increases
as the total number of genes increases.
Note the exponent on the equations in the
figure.
Leads to a proposed maximum genome size of
about 20,000 genes: where each non-regulatory
gene has its own regulatory gene
Horizontal Gene Transfer
• Defined as DNA transfer across species lines
– As opposed to vertical gene transfer: genes transmitted from parent to
offspring through chromosome replication and cell division.
– Once considered unusual or controversial, it is now obvious that HGT is a
frequent event with major effects on all prokaryotic genomes.
– HGT has made the definition of “species” difficult in prokaryotes.
• Pathogenicity islands: regions of up to 100 kbp, often near tRNA genes and
often containing multiple prophage insertions. They contain genes
needed for pathogenic behavior, such as toxins and type II secretion
systems.
• The classical three sexual processes in prokaryotes:
– Conjugation: direct transfer of DNA between two cells. Certain plasmids have
genes that cause conjugation.
– Transduction: transfer of DNA through a bacteriophage intermediate
• Gene transfer agents (GTAs) are defective bacteriophages that package and transfer
random pieces of the bacterial genome, without killing the host cells.
– Transformation: uptake of naked DNA from the environment.
More HGT
•
In the absence of a direct genome
comparison, horizontal gene transfer
can be detected by differences in DNA
composition: GC content, codon
usage, oligonucleotide frequency, etc.
– However, acquired genes undergo a
process of “amelioration”, where
selectively neutral mutations shift the
DNA composition to match the host’s
DNA.
•
Organisms that share a common
environment often transfer genes,
even across the bacteria-archaea
divide.
– Hyperthermophilic bacteria have up to
20% of their genes with better
matches in the archaea than in other
bacteria.
– Similarly, mesophilic archaea sahre
more genes with mesophilic bacteria
Top= bacteria, bottom = archaea.
In both cases, a mesophile is on the left
And a hyperthermophile is on the right.
HGT and Gene Loss
• Genes are gained by horizontal gene transfer as well as by internal
processes like duplication, and genes are also lost. The relative rates of
these two events must be balanced to keep the genome reasonably
constant in size.
• Probably all COG groups have had at least one horizontal transfer.
– But still, most genes are transferred vertically most of the time.
– Several studies have shown that most genes within a group of organisms have
a common phylogeny that matches the expectation of vertical descent.
• Are genes involved in replication, transcription, and translation less prone
to HGT? Based on the idea that these genes interact so intimately that
they can’t be easily replaced. However, many cases of HGT in these genes
have been seen, and there probably isn’t a big difference in rates.
– The problem is, much HGT in informational genes is not easily detected because the
COG families for these genes are “core”: nearly all species use slight variations on the
same gene. Genes involved in the same metabolic function tend to fall into several
different, non-homologous COGs.
Selfish Operons
• An operon can be thought of as a group of genes that act
together to perform a single metabolic function.
– Operons often even come with their own regulatory protein.
– An operon thus provides a phenotype that natural selection can act
on.
• You can think of operons as selfish: travelling between
genomes, conferring a useful trait that increases the number
of copies of that operon in the world.
• An example: membrane-bound ATP synthases, the primary
way most species generate energy. There is an archaeal
version and a bacterial version. Both are encoded by a single
operon, which has been transferred many times across the
domain boundary.
The Prokaryotic Mobilome
•
•
•
•
•
“-ome” means the set of all things with this function. I will admit to feeling that
this suffix is over-used these days.
The mobilome is the set of bacteriophages, plasmids, transposable elements, and
associated genes that frequently travel between the genomes of cellular life.
All sequenced genomes show signs of multiple integrated phages and plasmids
Bacteriophages are everywhere: it has been estimated that there at 10 times as
many phage particles as cells in sea water.
Plasmids are replicons independent of the chromosome.
– Usually circular, but sometimes linear.
– Usually not necessary for life, but some are very integrated into the life of the cell.
– Some integrate into the chromosome.
• Plasmid addiction: seen in restriction/modification systems and
toxin/antitoxin systems
– Put the toxin gene on the chromosome and the antitoxin on the plasmid. If the plasmid
is lost, the cell dies because it makes the toxin but not its antidote.
– This is a selfish process: the plasmid benefits but the cell is forced to keep replicating a
useless plasmid.
The Principal Processes of Prokaryotic Evolution
• Three basic processes:
– 1. vertical transfer from parent to child
– 2. horizontal transfer between species
– 3. mobilome genes that are occasionally recruited to perform useful functions
for the cells.
• Another important process: gene loss
– Sometimes under strong selection, as is the development of parasitism, where
many genes are no longer needed.
– More generally, there is a weak selection pressure to remove unneeded genes.
• The gene-centric perspective, as opposed to the classical genome-centric
viewpoint. Individual genes can be considered distinct evolutionary units
that are subject to selection across species and compete with other genes.
– In the gene-centric view, a genome is a community of genes that have some
degree of selfishness.
– Both views have validity.