Download Comparative mycobacterial genomics Stewart T Cole

567
Comparative mycobacterial genomics
Stewart T Cole
Genomics is providing us with a mass of information about the
biochemistry, physiology and pathogenesis of Mycobacterium
tuberculosis and Mycobacterium leprae. Comparison of the
two genome sequences is mutually enriching and indicates
that the M. leprae genome appears to have undergone
shrinkage and large-scale gene inactivation, which may
account for the exceptionally slow growth of this organism.
Addresses
Unité de Génétique Moléculaire Bactérienne, Institut Pasteur, 28 rue
du Docteur Roux, 75724 Paris Cedex 15, France;
e-mail: [email protected]
Current Opinion in Microbiology 1998, 1:567–571
http://biomednet.com/elecref/1369527400100567
© Current Biology Ltd ISSN 1369-5274
Abbreviations
C
cytosine
G
guanine
MPTR major polymorphic tandem repeat
PGRS polymorphic G + C rich sequence
Introduction
Intracellular pathogens such as the leprosy and tubercle
bacilli are amongst the most genetically intractable
microorganisms as a result of their long generation times,
fastidious growth requirements and contagiousness. Fresh
insight into the biology and pathogenicity of these important agents of human disease has been obtained by harnessing the powerful combination of genomics and
bioinformatics. The availability of the complete genome
sequence of Mycobacterium tuberculosis and much of the
sequence of the chromosome of Mycobacterium leprae has
resulted in a quantum leap in our knowledge and understanding. Comparison of the two datasets has identified
both common and specific genes that may be of relevance
to mycobacterial physiology and the respective diseases.
This review presents some of the highlights of mycobacterial genome analysis and illustrates how comparative
genomics may help us to understand phenotypic differences between related species.
The genome sequence of M. tuberculosis
H37Rv
The complete genome sequence of the well characterised
H37Rv strain of M. tuberculosis comprises 4,411,529 bp
and contains around 4000 genes accounting for >91% of
the potential coding capacity [1••,2•]. The high guanine
(G) and cytosine (C) content of the DNA (65.6%) affects
at least two other genomic parameters. First, the genome
contains a partial complement of tRNA genes as only 43
of the 61 possible tRNAs are encoded indicating that
extensive wobbling must occur during translation of
mRNAs. Second, the amino acid content of the proteome
is somewhat biased as amino acids encoded by G + C rich
codons are statistically more abundant than expected. To
identify regions with atypical base composition, that are
often indicative of horizontally transferred genes or coding sequences involved in pathogenesis, the G + C composition of the genome was plotted and found to be
relatively uniform. Several areas with exceptionally high
G + C content (>80%) were detected and shown to correspond to the PGRS (polymorphic G + C rich sequence)
gene family described below. Inspection of the regions
with higher than average adenine (A) and thymine (T)
content revealed genes encoding polyketide synthases or
transmembrane proteins (hydrophobic amino acids are
encoded by low G + C codons). Since these proteins are
probably involved in housekeeping functions this suggests that the acquisition of virulence genes in the form of
pathogenicity islands by horizontal transfer, as commonly
described in Gram-negative bacteria [3], may not have
occurred. The genome of H37Rv contains at least two
prophages, phiRv1 and phiRv2 [1••], and one of these was
probably lost during the attenuation process of M. bovis
that gave rise to bacille Calmette-Guérin (BCG) [4].
PhiRv1 is unlikely to contribute to pathogenesis because
it is not present in all clinical isolates of M. tuberculosis,
although no comparative studies of the degrees of virulence of isogenic strains with or without the prophage
have yet been performed.
Genes, their functions and diversity
Database comparisons (i.e. similarity of M. tuberculosis
genes to other genes of known function) have led to the
tentative attribution of functions to roughly 40% of the
3924 protein-coding genes identified and these are predominantly involved in core metabolism. Some functional information or similarity to other gene products
was found for a further 44% of the protein-coding genes,
although over half of these belong to the class known as
conserved hypotheticals, proteins of unknown function
but conserved sequence found in a variety of bacteria.
The remaining genes are probably characteristic of
mycobacteria as they show no similarity to any other
microbial sequences. Recent improvements in transposon mutagenesis and gene replacement technology in
mycobacteria [5•,6 •], coupled with more molecular
methods such as DNA chip technology, serial analysis of
gene expression and proteomics, should result in the
elucidation of further functions.
About 51% of the genes have arisen from gene duplication events. Although this value is close to that seen in
Escherichia coli and Bacillus subtilis [7,8] (i.e. eubacteria
with similar sized genomes), the degree of sequence conservation is much higher, suggesting that there may be
568
Genomics
extensive functional redundancy or that M. tuberculosis is
of recent evolutionary descent. Using systematic DNA
sequence analysis of 26 loci in a very large number of
independent strains, Musser and his colleagues have
demonstrated that there is a singular lack of sequence
diversity in the M. tuberculosis complex and conclude that
this is indicative of recent global dissemination [9••,10].
The basis for this remarkable genetic homogeneity is
unknown but most intriguing, and presumably reflects
either a very efficient DNA repair system or replication
machinery of exceptionally high fidelity. Inspection of
the genome for genes associated with these functions
provides some insight. First, M. tuberculosis may have an
exceptionally clean pool of nucleotide precursors because
it has three copies of mutT which encodes the enzyme
that removes the oxidised guanines whose incorporation
during replication causes base-pair mismatching. Second,
the genome does not appear to contain a mismatch repair
system because the key genes mutH, mutL, mutS and recJ
could not be found. Because mispaired bases are the substrate for the MutHLS system and may occur less frequently in M. tuberculosis than in other bacteria due to the
greater cleansing action of MutT, the mismatch repair
system may not be required.
The predominant gene families
More than 20% of the M. tuberculosis chromosome is
devoted to genes encoding two different classes of proteins: enzymes involved in fatty acid metabolism and
acidic, glycine-rich polypeptides of unknown function,
the PE and PPE proteins [1••,11]. The mycobacterial cell
envelope contains a dazzling array of lipids, glycolipids,
mycolic acids and polyketides [12,13], and it was thus no
great surprise to find encoded in the genome examples of
every known lipid and polyketide biosynthetic system,
including enzymes usually confined to mammals and
plants, in addition to the more common bacterial systems.
More genes encoding potential lipid biosynthetic activities were uncovered than there are known metabolites
thus raising the intriguing possibility that several novel
lipid and polyketide species may exist [11]. It is conceivable that these unknown metabolites may only be produced in restricted environments such as the macrophage
and that some of them may even have immunomodulatory activity like the potent polyketide, rapamycin, and
account for the pronounced immune suppression
observed in mycobacterial diseases.
Although there are extensive lipid biosynthetic functions
in M. tuberculosis, these are eclipsed by the vast selection
of genes and enzymes potentially involved in fatty acid
degradation. In addition to the multifunctional
FadA/FadB proteins that effect all the reactions of the βoxidation cycle, there are >100 genes encoding enzymes
that could catalyse the individual steps. These may well
be involved in the degradation of lipids present in host
membranes and could make important contributions to
energy metabolism [14].
One of the great surprises of the M. tuberculosis genome
project was the discovery of two large gene families encoding unusual glycine-rich proteins with basic pIs (isoelectric
points) and well conserved amino-terminal domains.
These show no significant similarity to proteins of known
function and were referred to as the PE and PPE protein
families as they contain conserved Pro–Glu and
Pro–Pro–Glu motifs at the amino terminus, respectively
[1••,11]. There are 99 members of the PE family and 61 of
these belong to the PGRS subfamily, containing multiple
tandem repetitions of the tripeptide Gly–Gly–Ala (or a
variant thereof), whereas the remaining PE proteins have
different carboxy-terminal domains. The PPE family comprises 68 members that fall into at least three subfamilies,
the most spectacular of which is the MPTR (major polymorphic tandem repeat) class as several of these proteins
are predicted to consist of >3000 amino acid residues mainly repetitions of the motif Asn-X-Gly-X-Gly-Asn-X-Gly
(where X is any amino acid). The PGRS members of the
PE subfamily are also large proteins and can contain up to
1400 amino acid residues. As the names PGRS and MPTR
imply, the genes based on these sequences show extensive
polymorphisms in the different M. tuberculosis complex
strains probably as a result of strand slippage during replication of the many simple sequences comprising the coding sequence [15–17]. Consequently, the corresponding
proteins should also display size and sequence variation
and it is possible that they may even represent variable
protein antigens that are relevant in evasion of the host
immune responses.
The genome sequence of M. leprae
Unlike M. tuberculosis, there were no alternative genetic
approaches to tackle the genetics of the leprosy bacillus
and systematic mapping and sequencing was thus initiated at an early stage [18]. At the present time close to
90% of the genome has been sequenced and much of
this has been analysed and annotated [19,20]. On inspection of the first cosmid sequence [21] it was apparent
that the gene density at ~50–60% was unusually low for
a bacterium, and this trend has since been observed
throughout the genome although there are regional differences [19]. Now that the complete genome sequence
of M. tuberculosis H37Rv is available [1••] detailed comparisons can be undertaken and these are proving to be
extremely useful in understanding the basis of the biology of these important pathogens. As first indicated by
hybridisation mapping [22], the two genomes show synteny but this appears to be local rather than global. In
most instances, the same genes or operons occur in the
same order but these are often connected by extended
regions that are apparently noncoding or contain pseudogenes in M. leprae.
An example of these noncoding regions or pseudogenes is
shown in Figure 1, where a segment situated between the
rif and str operons of both M. tuberculosis and M. leprae is
aligned, the genes identified and differences highlighted. It
Comparative mycobacterial genomics Cole
is immediately clear that in M. tuberculosis this region is both
larger and contains more genes than the corresponding
stretch from M. leprae. If one makes the reasonable assumption that both mycobacteria are descended from a common
ancestor then they should have similar sized genomes. This
raises the possibility that either the leprosy bacillus has lost
genes or that M. tuberculosis has acquired additional functions as a result of gene duplication or horizontal transfer.
Dotplot analysis of the corresponding nucleotide sequences
provides support for both interpretations.
569
Figure 1
end
rpsL
end
lpqP
fadE8
echA4
Rv0674
Of the 13 genes found in M. tuberculosis in this region
(Figure 1), only two have been conserved in M. leprae and
these show between 80–90% identity with their counterparts. Lower levels of sequence identity (~70%) can be
seen in regions equivalent to six M. tuberculosis coding
sequences; in the leprosy bacillus the corresponding
regions contain pseudogenes with multiple frameshifts,
small deletions and in-frame stop codons which would
abolish gene expression. These data suggest that functional genes were once present in M. leprae but that they
have been silenced because their activities were no
longer required by an obligate intracellular parasite.
Three of these genes (fadE8, echA4, echA5) encode putative β-oxidation enzymes that could degrade unknown
fatty acids thereby indicating that M. tuberculosis has the
potential to metabolise a larger choice of substrates for
growth than M. leprae. Functional information is available
for a further three genes: lpqP encodes a lipoprotein that
is missing from M. leprae, whereas the other two mmpL5
and mmpS5 encode conserved integral membrane proteins that are confined to mycobacteria and may effect
specific tasks such as metabolite transport. As the latter
genes belong to a family, it is probable that they arose via
a duplication event.
The presence of large numbers of pseudogenes in M. leprae probably accounts for many of the phenotypic differences between the leprosy and tubercle bacilli and this is
exemplified by their respective responses to the key tuberculocidal agent, isoniazid. In M. tuberculosis, katG encodes
catalase-peroxidase, a heme-containing enzyme that mediates the toxic effect of the key tuberculocidal agent, isoniazid [23]. Comparison of the corresponding regions of
M. leprae and M. tuberculosis revealed the presence of
numerous mutations in the M. leprae gene that abolished
its activity [24,25]. This undoubtedly explains why the
leprosy bacillus produces no catalase-peroxidase and displays high level resistance to isoniazid.
Until the genome sequence of M. leprae is completed, the
precise number of genes will remain unknown.
Preliminary estimates suggest that the proteome may contain as few as 1600 proteins. The genome of M. leprae is
roughly 1.4 Mb smaller than that of M. tuberculosis and its
G + C content (57%) is significantly lower than those of all
other mycobacterial genomes. Although deletion of coding
sequences almost certainly led to some genome shrinkage,
echA5
mmpL5
mmpS5
Rv0678
Rv0679c
Rv0680c
Rv0681
rpsL
Current Opinion in Microbiology
Dotplot alignment of the nucleotide sequences found in the region
downstream of the rif operons of M. leprae (horizontal axis) and
M. tuberculosis (vertical axis). The M. leprae sequence is 9,186 bp in
length and that of M. tuberculosis 12,140. Both sequences start with
the initiation codon of the end gene (endonuclease IV) and finish with
the termination codon of rpsL (r-protein S12). M. tuberculosis has an
additional lipoprotein gene, lpqP, between end and fadE8, and an
operon mmpL5mmpS5 that is not found in M. leprae. By contrast,
M. leprae appears to have additional DNA in the region between the
residual Rv0681 sequences and rpsL. The figure was generated by
DOTTER [27] using a window of 25 bases and greyramp settings of
41–101. Where the degree of relatedness of the two sequences is
above the average score, a dot is plotted, and its intensity is
proportional to its score. When the dots merge to form a line, this
indicates that the DNA sequences are highly related and correspond
to genes or pseudogenes. See the text for a detailed explanation of
similarity values and further biological interpretation.
two other factors may also have contributed to the size difference. First, M. leprae contains very few members of the
PE and PPE gene families which account for ~450 kb of
the M. tuberculosis chromosome [1••]. Second, traces of far
fewer insertion sequences (IS) and bacteriophages have
been found in M. leprae than in M. tuberculosis H37Rv,
where they contribute over 120 kb.
Comparisons such as those outlined above with M. tuberculosis will be extremely informative as they will allow the
genes and their functions to be classed into three broad
groups: those found in many bacteria, those confined to the
genus Mycobacterium and those present only in one or other
species. It is already clear that a small number of proteins
570
Genomics
have no counterparts in M. tuberculosis and these may confer novel biological properties on M. leprae, and be involved
in functions such as neurotropism or nerve damage [26••].
Conclusions
Genomics will radically change the way mycobacteriologists
tackle problems of pathogenesis by enabling more focused
experimental approaches to be adopted and providing
greater subject diversity. Comparative genomics will lead to
the identification of genes restricted to a given mycobacterium that may play unique biological roles, and serve as
sources of specific antigens or potential drug targets.
Acknowledgements
complete genome sequence of Escherichia coli K-12. Science
1997, 277:1453-1462.
8.
Kunst F, Ogasawara N, Moszer I, Albertini AM, Alloni G, Azevedo V,
Bertero MG, Bessieres P, Bolotin A, Borchert S et al.: The complete
genome sequence of the Gram-positive bacterium Bacillus
subtilis. Nature 1997, 390:249-256.
9.
••
Sreevatsan S, Pan X, Stockbauer KE, Connell ND, Kreiswirth BN,
Whittam TS, Musser JM: Restricted structural gene polymorphism
in the Mycobacterium tuberculosis complex indicates
evolutionarily recent global dissemination. Proc Natl Acad Sci
USA 1997, 94:9869-9874.
This highly interesting study presents the results of a rigorous attempt to
understand the population genetics of the members of the M. tuberculosis
complex and demonstrates that there is very little genetic drift in the set of
sequences examined. The findings are interpreted in an evolutionary context
and suggest that tuberculosis may be a very recent disease of humans.
10. Kapur V, Whittam TS, Musser J: Is Mycobacterium tuberculosis
15,000 years old? J Infect Dis 1994, 170:1348-1349.
I gratefully acknowledge support provided by the Wellcome Trust, the
Association Française Raoul Follereau, the World Health Organisation and
the Institut Pasteur.
11. Cole ST, Barrell BG: Analysis of the genome of Mycobacterium
tuberculosis H37Rv. In Genetics and Tuberculosis (Novartis
Foundation Symposium 217). Edited by Chadwick DJ, Cardew G.
Chichester: John Wiley; 1998:160-172.
References and recommended reading
12. Kolattukudy PE, Fernandes ND, Azad AK, Fitzmaurice AM,
Sirakova TD: Biochemistry and molecular genetics of cell-wall lipid
biosynthesis in mycobacteria. Mol Microbiol 1997, 24:263-270.
Papers of particular interest, published within the annual period of review,
have been highlighted as:
• of special interest
•• of outstanding interest
1.
••
Cole ST, Brosch R, Parkhill J, Garnier T, Churcher C, Harris D,
Gordon SV, Eiglmeier K, Gas S, Barry CE III et al.: Deciphering the
biology of Mycobacterium tuberculosis from the complete
genome sequence. Nature 1998, 393:537-544.
The authors describe the complete genome sequence of the most widely used
strain of M. tuberculosis H37Rv and interpret the genomic message using
various bioinformatic methods. The main themes addressed are slow growth,
dormancy, genetic homogeneity, biosynthesis of the cell envelope and the PE
and PPE gene families. The possibility that PE and PPE proteins may be
sources of antigenic variation or inhibitors of antigen presentation is discussed.
2.
•
Brosch R, Gordon SV, Billault A, Garnier T, Eiglmeier K, Soravito C,
Barrell BG, Cole ST: Use of a Mycobacterium tuberculosis H37Rv
Bacterial Artificial Chromosome (BAC) library for genome
mapping, sequencing and comparative genomics. Infect Immun
1998, 66:2221-2229.
This paper describes the first successful application of BACs in bacterial
genomics and shows how complete genome coverage could be obtained in
contrast to the situation with cosmids. The complete genome can be
represented by 68 BACs. An application in comparative genomics is
presented and a 12.7 kb region of difference between the chromosomes of
M. tuberculosis, M. bovis and M. bovis BCG characterised.
3.
Hacker J, Blum OG, Muhldorfer I, Tschape H: Pathogenicity islands
of virulent bacteria: structure, function and impact. Mol Microbiol
1997, 23:1089-1097.
4.
Mahairas GG, Sabo PJ, Hickey MJ, Singh DC, Stover CK: Molecular
analysis of genetic differences between Mycobacterium bovis
BCG and virulent M. bovis. J Bacteriol 1996, 178:1274-1282.
5.
•
Pelicic V, Jackson M, Reyrat JM, Jacobs WR Jr, Gicquel B, Guilhot C:
Efficient allelic exchange and transposon mutagenesis in
Mycobacterium tuberculosis. Proc Natl Acad Sci USA 1997,
94:10955-10960.
This publication describes the development of a highly efficient vector
system that greatly facilitates gene replacements in mycobacteria. This
system is certain to find widespread application. Preliminary results of its
use as a vehicle for transposon delivery are presented.
6.
•
Bardarov S, Kriakov J, Carriere C, Yu S, Vaamonde C, McAdam R,
Bloom BR, Hatfull GR, Jacobs JWR: Conditionally replicating
mycobacteriophages: a system for transposon delivery to
Mycobacterium tuberculosis. Proc Natl Acad Sci USA 1997,
94:10961-10966.
The authors describe the construction of a synthetic transposon and its
application as an insertional mutagen in M. tuberculosis. The results are
interpreted in the context of the genome sequence and suggest that this
approach could be very useful for identifying and inactivating nonessential
genes. Some bias in the sites of insertion was seen that reflects the locations
of the endogenous IS (insertion sequence) elements in the genome.
7.
Blattner FR, Plunkett G III, Bloch CA, Perna NT, Burland V, Riley M,
Collado-Vides J, Glasner JD, Rode CK, Mayhew GF et al.: The
13. Besra GS, Chaterjee D: Lipids and carbohydrates of
Mycobacterium tuberculosis. In Tuberculosis: Pathogenesis,
Protection, and Control. Edited by Bloom BR. Washington DC:
American Society for Microbiology; 1994:285-306.
14. Wheeler PR, Ratledge C: Metabolism of Mycobacterium
tuberculosis. In Tuberculosis: Pathogenesis, Protection, and Control.
Edited by Bloon BR. Washington DC: American Society for
Microbiology; 1994:353-385.
15. Hermans PWM, van Soolingen D, van Embden JDA: Characterization
of a major polymorphic tandem repeat in Mycobacterium
tuberculosis and its potential use in the epidemiology of
Mycobacterium kansasii and Mycobacterium gordonae. J Bacteriol
1992, 174:4157-4165.
16. Poulet S, Cole ST: Characterisation of the polymorphic GC-rich
repetitive sequence (PGRS) present in Mycobacterium
tuberculosis. Arch Microbiol 1995, 163:87-95.
17.
Poulet S, Cole ST: Repeated DNA sequences in mycobacteria.
Arch Microbiol 1995, 163:79-86.
18. Eiglmeier K, Honoré N, Woods SA, Caudron B, Cole ST: Use of an
ordered cosmid library to deduce the genomic organisation of
Mycobacterium leprae. Mol Microbiol 1993, 7:197-206.
19. Smith DR, Richterich P, Rubenfield M, Rice PW, Butler C, Lee H-M,
Kirst S, Gundersen K, Abendschan K, Xu Q et al.: Multiplex
sequencing of 1.5 Mb of the Mycobacterium leprae genome.
Genome Res 1997, 7:802-819.
20. Fsihi H, De Rossi E, Salazar L, Cantoni R, Labò M, Riccardi G,
Takiff HE, Eiglmeier K, Bergh S, Cole ST: Gene arrangement and
organisation in a ~76 kilobase fragment encompassing the oriC
region of the chromosome of Mycobacterium leprae. Microbiology
1996, 142:3147-3161.
21. Honoré N, Bergh S, Chanteau S, Doucet-Populaire F, Eiglmeier K,
Garnier T, Georges C, Launois P, Limpaiboon P, Newton S et al.:
Nucleotide sequence of the first cosmid from the Mycobacterium
leprae genome project: structure and function of the Rif-Str
regions. Mol Microbiol 1993, 7:207-214.
22. Philipp WJ, Poulet S, Eiglmeier K, Pascopella L, Subramanian B,
Heym B, Bergh S, Bloom BR, Jacobs WR Jr, Cole ST: An integrated
map of the genome of the tubercle bacillus, Mycobacterium
tuberculosis H37Rv, and comparison with Mycobacterium leprae.
Proc Natl Acad Sci USA 1996, 93:3132-3137.
23. Zhang Y, Heym B, Allen B, Young D, Cole S: The catalaseperoxidase gene and isoniazid resistance of Mycobacterium
tuberculosis. Nature 1992, 358:591-593.
24. Eiglmeier K, Fsihi H, Heym B, Cole ST: On the catalase-peroxidase
gene, katG, of Mycobacterium leprae and the implications for
treatment of leprosy with isoniazid. FEMS Microbiol Lett 1997,
149:273-278.
Comparative mycobacterial genomics Cole
25. Nakata N, Matsuoaka M, Kashiwabara Y, Okada N, Sasakawa C:
Nucleotide sequence of the Mycobacterium leprae katG region.
J Bacteriol 1997, 179:3053-3057.
26. Rambukkana A, Salzer JL, Yurchenco PD, Tuomanen EI: Neural
•• targeting of Mycobacterium leprae mediated by the G domain of
the laminin-a2 chain. Cell 1997, 88:811-821.
This is the most significant publication in the past decade addressing the
tropism and mechanism of cell invasion of M. leprae. The authors
demonstrate that the bacillus interacts with the G domain of the laminin
571
alpha 2 chain and that this is necessary and sufficient for adherence to
Schwann cells. This protein probably acts as a tissue-restricted bridging
molecule and would explain why the leprosy bacillus is found in peripheral
nerves and muscle cells. It should now be possible to identify the ligand on
the bacterial surface that interacts with the G domain.
27.
Sonnhammer ELL, Durbin R: A dot-matrix program with dynamic
threshhold control suitable for genomic DNA and protein
sequence analysis. Gene 1995, 167:GC1-10.