Comparative Genomic Analysis of Fruiting Body Formation in
Myxococcales
Stuart Huntley,1 Nils Hamann,1 Sigrun Wegener-Feldbrügge,1 Anke Treuner-Lange,1 Michael Kube,2
Richard Reinhardt,2 Sven Klages,2 Rolf Müller,3 Catherine M. Ronning,4 William C. Nierman,4 and
Lotte Søgaard-Andersen*,1
1
Department of Ecophysiology, Max Planck Institute for Terrestrial Microbiology, Marburg, Germany
Max Planck Institute for Molecular Genetics, Berlin, Germany
3
Department Microbial Natural Products, Helmholtz Institute for Pharmaceutical Research Saarland, Saarland University,
Saarbrücken, Germany
4
Infectious Diseases Program, J. Craig Venter Institute, Rockville, MD
*Corresponding author: E-mail: [email protected].
Associate Editor: James McInerney
2
Genetic programs underlying multicellular morphogenesis and cellular differentiation are most often associated with
eukaryotic organisms, but examples also exist in bacteria such as the formation of multicellular, spore-filled fruiting bodies
in the order Myxococcales. Most members of the Myxococcales undergo a multicellular developmental program
culminating in the formation of spore-filled fruiting bodies in response to starvation. To gain insight into the evolutionary
history of fruiting body formation in Myxococcales, we performed a comparative analysis of the genomes and
transcriptomes of five Myxococcales species, four of these undergo fruiting body formation (Myxococcus xanthus,
Stigmatella aurantiaca, Sorangium cellulosum, and Haliangium ochraceum) and one does not (Anaeromyxobacter
dehalogenans). Our analyses show that a set of 95 known M. xanthus development-specific genes—although suffering from
a sampling bias—are overrepresented and occur more frequently than an average M. xanthus gene in S. aurantiaca,
whereas they occur at the same frequency as an average M. xanthus gene in S. cellulosum and in H. ochraceum and are
underrepresented in A. dehalogenans. Moreover, genes for entire signal transduction pathways important for fruiting body
formation in M. xanthus are conserved in S. aurantiaca, whereas only a minority of these genes are conserved in
A. dehalogenans, S. cellulosum, and H. ochraceum. Likewise, global gene expression profiling of developmentally regulated
genes showed that genes that upregulated during development in M. xanthus are overrepresented in S. aurantiaca and
slightly underrepresented in A. dehalogenans, S. cellulosum, and H. ochraceum. These comparative analyses strongly
indicate that the genetic programs for fruiting body formation in M. xanthus and S. aurantiaca are highly similar and
significantly different from the genetic program directing fruiting body formation in S. cellulosum and H. ochraceum. Thus,
our analyses reveal an unexpected level of plasticity in the genetic programs for fruiting body formation in the
Myxococcales and strongly suggest that the genetic program underlying fruiting body formation in different Myxococcales
is not conserved. The evolutionary implications of this finding are discussed.
Key words: Sporulation, fruiting body formation, bacterial development, bacterial differentiation, Myxococcus xanthus,
comparative genomics.
Introduction
To optimize their chances of survival in an ever-changing
environment, bacteria have several strategies to sense and
respond to these changes (Perez and Groisman 2009). One
strategy involves adaptive changes in gene expression without evident changes in cell morphology. In an alternative
strategy, changes in gene expression result in cellular differentiation with the formation of cells with altered morphologies and properties. One of the best-studied examples of
cell differentiation in bacteria in response to changes in the
environment is the formation of a heat and desiccationresistant cell morphology, referred to as spores and cysts
(from here on referred to collectively as spores). Spore
formation is generally induced by nutrient depletion and
spores are generally less metabolically active than vegetative
cells, are more resistant to various types of chemical and
physical stresses than vegetative cells, and can survive extended periods of starvation (Shimkets and Brun 2000).
Bacterial species from widely separated taxonomic clades
have evolved the ability to form spores in response to starvation (Shimkets and Brun 2000), suggesting that the ability
to form spores has emerged independently several times
in bacteria. Spore-forming bacteria include members of
the low G þ C Gram-positive bacteria (e.g., Bacillus subtilis
and Clostridium tetani), high G þ C Gram-positive bacteria
(e.g., Streptomyces coelicolor), filamentous cyanobacteria (e.g.,
Nostoc punctiforme) as well as members of the Gramnegative Alphaproteobacteria (e.g., Methylosinus trichosporium and Rhodospirillum capsulatus), Gammaproteobacteria
(e.g., Azotobacter vinelandii) and Deltaproteobacteria (e.g.,
© The Author 2010. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please
e-mail: [email protected]
Mol. Biol. Evol. 28(2):1083–1097. 2011 doi:10.1093/molbev/msq292
Advance Access publication October 29, 2010
1083
Research article
Abstract
MBE
Huntley et al. · doi:10.1093/molbev/msq292
Myxococcus xanthus) (Shimkets and Brun 2000). Spore formation has been most thoroughly studied in B. subtilis,
S. coelicolor, and M. xanthus. Supporting the idea that sporulation has evolved several times independently in bacteria,
the mechanisms underlying spore formation in these three
phylogenetically widely separated species are very different
(Shimkets and Brun 2000). In B. subtilis (Kroos 2007), spore
formation is initiated by an asymmetric cell division in
a rod-shaped cell resulting in the formation of a large
mother cell and a small prespore. The prespore is subsequently engulfed by the mother cell before the mother cell
lyses resulting in the release of the spherical mature spore.
In S. coelicolor (Flärdh and Buttner 2009), chains of spherical spores are formed by synchronous cell divisions in aerial
hyphae. In M. xanthus, spore formation is coupled to the
formation of multicellular fruiting bodies, and individual
rod-shaped cells that have accumulated inside a fruiting
body directly differentiate into spherical spores without
a prior cell division (Kroos 2007). The availability of complete genome sequences of related bacteria provide the opportunity to explore to what extent the genetic programs
governing spore formation in a taxonomically well-defined
group of bacteria are conserved. Here, we have focused on
an evolutionary analysis of the genetic program underlying
spore formation in the Myxococcales.
Members of the Deltaproteobacteria order Myxococcales are typically found in topsoil where they form cooperatively feeding colonies and grow as saprophytes by
decomposing degradable polymers or as predators by preying on other microorganisms (Shimkets et al. 2006). They
share three traits that distinguish them from other Deltaproteobacteria. First, most members are strictly aerobic,
whereas most Deltaproteobacteria are anaerobic (Shimkets
et al. 2006). Second, in response to starvation, most Myxococcales initiate a developmental program that culminates in the formation of multicellular fruiting bodies
inside which the rod-shaped cells differentiate into spores
(Shimkets et al. 2006). Third, most members of the Myxococcales have large genome sizes around 10 Mb (Goldman
et al. 2006; Schneiker et al. 2007; Thomas et al. 2008; Wu
et al. 2009; Natalia et al. 2010), whereas other Deltaproteobacteria have smaller genome sizes ranging from 2.9 to 6.5
Mb (supplementary table S1, Supplementary Material
online).
Myxococcus xanthus has emerged as the model organism
to understand the molecular mechanisms underlying fruiting body formation (Kroos 2007). Fruiting body formation
is initiated by the RelA-dependent stringent response and
proceeds in distinct morphological stages. The first signs of
fruiting body formation are evident 4–6 h after starvation
as cells aggregate to form small aggregation centers. As they
accumulate more cells, the centers increase in size and
eventually become mound shaped. By 24 h, the aggregation
process is complete and each nascent fruiting body contains approximately 105 cells. Inside the nascent fruiting
bodies, the rod-shaped cells differentiate into spherical
spores, resulting in mature fruiting bodies. Of the total
population present at the onset of starvation, only 10%
1084
undergo sporulation and normally only those cells that
have accumulated inside the fruiting bodies. Up to 30%
of the cells remain outside the fruiting bodies. These cells
remain rod shaped and differentiate to a cell type termed
peripheral rods (O’Connor and Zusman 1991a, 1991b). Finally, the remaining cells undergo lysis (Wireman and
Dworkin 1977; Rosenbluh et al. 1989). Recently, this cell lysis
was suggested to reflect programmed cell death (Nariya
and Inouye 2008). Fruiting body formation depends on intercellular signaling (Kroos 2007) and highly coordinated
changes in gene expression (Kroos 2007), as well as on
the coordinated motility behavior of cells (Leonardy
et al. 2008). Thus, fruiting body formation includes many
of the characteristics associated with developmental programs in metazoans such as cell–cell signaling, cells adopting different cell fates, programmed cell death,
morphogenetic cell movements, and coupling of cell differentiation to multicellular morphogenesis. Because spore
formation occurs inside the fruiting bodies, these structures
are essentially spore-filled assemblies. In the presence of nutrients, spores germinate to give rise to rod-shaped, motile,
metabolically active vegetative cells. With each fruiting
body consisting of approximately 105 spores, germination
immediately gives rise to a spreading, cooperatively feeding
colony. Thus, it has been argued that fruiting bodies
are optimally designed to ensure that a new vegetative
cycle is initiated by a community of cells rather than by
a single cell (Dworkin 1972).
All Myxococcales tested—with the exception of one (see
below)—initiate fruiting body formation in response to
starvation, suggesting that the last common ancestor of
the Myxococcales harbored a genetic program for fruiting
body formation and that fruiting Myxococcales would
share in common a genetic program underlying fruiting
body formation. Here, we examined this hypothesis using
comparative and functional genomics on four complete genomes of fruiting Myxococcales and one genome of the
only known nonfruiting Myxococcales species. Our analyses suggest that the genetic program underlying fruiting
body formation in different Myxococcales is not conserved,
that is, different species of Myxococcales harbor different
programs for fruiting body formation. The evolutionary
implications of this finding are discussed.
Materials and Methods
Phylogenetic Analyses
Full-length 16S rRNA sequences of type strains representing diverse Myxococcales genera, as well as sequences from
several non-Myxococcales Proteobacteria were collected
from the Ribosomal Database Project (http://rdp.cme.msu.edu/). The sequences were aligned using ClustalW1.83
(Thompson et al. 1994), ends of the resulting alignment
were trimmed so that all aligned sequences extended fully
to both ends of the alignment. Gaps were removed and the
sequences were realigned using ClustalW1.83 to ensure
that the modifications from the first alignment did not alter the relationships between the remaining sequences. The
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
FIG. 1. Phylogeny within the order Myxococcales and fruiting body
morphologies. Dendrogram of members of the order Myxococcales.
Species included in this study are indicated by the asterisks (*). A
consensus sequence generated from an alignment of several nonMyxococcales 16S rRNA genes was used as the root of the tree.
Bootstrap values (percentages) shown at nodes are based on 1,000
replications. The vertical bar to the right indicates the three
suborders within the Myxococcales: light gray, Cystobacterineae;
dark gray, Nannocystineae; and black, Sorangineae.
resulting alignment was used to generate a bootstrapped
(1,000 iterations) neighbor joining tree file from which the
image shown in Figure 1 was generated using the Treedyn
software (Chevenet et al. 2006). The above analyses were
repeated using the 16S rRNA sequences from all Deltaproteobacteria with a completed genome sequence to produce the tree figure shown in supplementary table S1
(Supplementary Material online).
Sequencing and Annotation of Stigmatella
aurantiaca DW4/3-1
Stigmatella aurantiaca DW4/3-1 DNA was isolated with
the Genomic DNA kit (Qiagen, Hildesheim, Germany) according to the manufacturer’s instructions. The genomic
sequences were determined by whole-genome shotgun sequencing using Sanger-based sequencing technology and
pyrosequencing. The genome was covered by short-insert
shotgun libraries and a fosmid library with 37-kb inserts
(CopyControl Fosmid Library Production Kit; Epicentre,
Madison). End sequencing of clone-derived inserts was performed using BigDye 3.1 chemistry and 3730XL capillary
sequencers (ABI, Darmstadt, Germany) resulting in
90,783 reads (plasmid-derived reads 61,770 and 27,683 fosmid-derived reads) reaching a 7-fold sequencing coverage.
In addition, pyrosequencing was performed using the GS20
sequencer (454 life science/Roche) resulting in 292,874
reads (241 bp average read length), which provided an additional 7-fold sequencing coverage. The assembly of
Sanger and 454 derived reads was performed within two
steps. GS20 data were initially assembled by Newbler
(454 life science/Roche) and the resulting contigs were fragmented using PERL scripts resulting in overlapping sequence fragments, which were assigned as forward and
reverse reads in fasta format and the corresponding fasta
MBE
quality files. These ‘‘faked reads’’ were assembled together
with the Sanger-derived processed reads using PhredPhrap
(http://www.phrap.org). GS FLX data and Sanger-derived
reads were assembled using the Celera assembler v4.3
(Myers et al. 2000). Assembled data were imported into
Consed (Gordon 2003), edited, and verified. Finishing experiments included primer walking on bridging fosmid
clones and polymerase chain reaction products to improve
sequence quality and gap closure to a single closed circular
sequence. Stigmatella aurantiaca protein-coding sequences
were predicted by Glimmer3 (Delcher et al. 1999), manually
curated in Artemis (Rutherford et al. 2000), and annotated
with HTGA (Rabus et al. 2002). Structural rRNAs and
tRNAs were determined using RNAmmer (Lagesen et al.
2007) and tRNAscan-SE (Lowe and Eddy 1997). An unfinished version of the S. aurantiaca DW4/3-1 genome was
previously reported (NZ_AAMD00000000). The genome
sequence reported here represents the complete genome
sequence of the DW4/3-1 strain. The genome sequence is
deposited in GenBank with the accession number
CP002271.
Functional Annotation and Ortholog Identification
Sequences and annotation for all publicly available bacterial species for which complete genome data available were
obtained from the NCBI site ftp://ftp.ncbi.nih.gov/genomes/Bacteria/, files ‘‘All.faa.tar.gz and All.ptt.tar.gz (6
May 2009). Based on our information concerning known
developmentally important genes, we included two M. xanthus loci (MXAN_3217, actD; and MXAN_5430, desS) to
those included in the publicly available annotation. Functional annotations used in this study were generated by
adaptation of available cluster of orthologous group
(COG) functional assignments. First, locus descriptions
from the four annotations were combined and sorted
based on product description. All loci bearing the same description were assigned a single functional assignment
code, chosen based on the code most prevalent among
those provided in the COG assignments. Loci for which
no COG assignment were available, as well as those loci
for which the ambiguous COG categories R (general prediction only) and S (conserved) were originally assigned,
were combined into a new category ‘‘X,’’ to indicate that
no clear function has been identified for these proteins. Following the reciprocal Blast analyses (see below), final adjustments to the functional assignments were made
manually such that orthologs shared the same functional
category assignment.
To identify orthologs and homologs for each Myxococcales protein, complete, independent protein sequence databases were generated for each of the Myxococcales
genome, as well as each of 973 complete, non-Myxococcales bacterial genomes available at the time of the analyses
(from NCBI data files listed above). Each Myxococcales protein was used as a query sequence in a BlastP analysis
against each database to identify homologs of the protein,
using an expect score cutoff of 1 105. The BlastPidentified matches were filtered, saving only those
1085
MBE
Huntley et al. · doi:10.1093/molbev/msq292
Table 1. Characteristics of the Five Myxococcales Species Analyzed in This Study.
Characteristic
Suborder
Genome size (bp)a
Protein-coding
sequencesa
Genome GC
contenta (%)
Metabolisma
Oxygen
tolerance/requirementb
Fruiting body
morphologyb
Spore morphologyb
Myxococcus
xanthus
Cystobacterineae
9,139,732
7,333
Stigmatella
aurantiaca
Cystobacterineae
10,260,755
8,352
Anaeromyxobacter
dehalogenans
Cystobacterineae
5,013,478
4,346
Sorangium
cellulosum
Sorangineae
13,033,778
9,384
Haliangium
ochraceum
Nannocystineae
9,446,314
6,719
68.9
Amino acids;
respiration
67.5
Amino acids;
respiration
74.9
Short fatty acids;
respiration
71.4
Cellulose;
respiration
69.0
DNA, amino acids,
starch; respiration
Aerobic
Hay-stack
shaped
Aerobic
Sporangioles
on stalk
Microaerophilic
Not fruiting
Aerobic
Spherical,
optically
refractile
Short bent rods,
optically refractile
NA
Aerobic
Assembly of
sporangioles,
not on stalk
Short rods,
optically refractile
Assembly of sporangioles,
not on stalk
Spherical, optically
refractile
NOTE.—NA, not applicable.
a
References for genome information: M. xanthus (Goldman et al. 2007), S. aurantiaca (this study), A. dehalogenans (Thomas et al. 2008), S. cellulosum (Schneiker et al.
2007), and H. ochraceum (Natalia et al. 2010).
b
References for physiology, fruiting body, and spore morphology: M. xanthus, S. aurantiaca, and S. cellulosum (Shimkets et al. 2006), A. dehalogenans (Sanford et al. 2002),
and H. ochraceum (Fudou et al. 2002).
matches that spanned at least one-third of the query sequence length. The expect scores used and the alignment
length filter size were empirically determined to produce
BlastP results with a maximum number of queries producing
significant hits while limiting random noise in the results. For
the purposes of this study, we considered all BlastP hits that
passed these requirements to be homologs of the initial
query protein. The protein corresponding to the highest
scoring (bit score) match was then used as a query in a second BlastP analysis, this time against a complete protein database for the organism from which the original query
protein was found. If the highest scoring match in this second BlastP analysis was to the original query protein, the two
proteins were considered orthologs for the purposes of this
study. To estimate the false-negative discovery rate of orthologs, we determined the frequency with which the original
query protein was among the five best hits but not the best
scoring hit in the second (reciprocal) BlastP analysis.
Synteny Plots
To visualize the conservation of genomic context of orthologs between species, a Perl script was written, which reads
the genomic coordinates of each pair of orthologs from
their respective genomes. Subsequently, the coordinates
of all pairs of orthologs were plotted in a diagram having
the coordinates of one species on the X axis and the coordinates of the other species on the Y axis giving rise
to a dot plot-style image.
Results
Comparisons of Five Completely Sequenced
Myxococcales Genomes
At the time of this study, complete genome sequence information was available for M. xanthus strain DK1622
(Goldman et al. 2006), Sorangium cellulosum strain So
1086
ce56 (Schneiker et al. 2007), Haliangium ochraceum strain
SMP-2 (Natalia et al. 2010), Anaeromyxobacter dehalogenans strains 2CP-C (Thomas et al. 2008) and 2CP-1,
and Anaeromyxobacter species 109-5 and K. To decide
which Anaerobacter genome to use in our analyses, we
compared the four Anaerobacter genomes. The genomes
of Anaeromyxobacter 2CP-C, 2CP-1, Fw109-5, and K contain 4,346, 4,473, 4,466, and 4,457 protein-coding sequences, respectively (supplementary fig. S1, Supplementary
Material online). Possible orthologs for each Anaeromyxobacter gene were identified using a reciprocal best BlastP
hit method (see below). About 3,041 genes are conserved
in all four genomes and represent 70%, 68%, 68%, and 68%
of the genes in 2CP-C, 2CP-1, Fw109-5, and K, respectively
(supplementary fig. S1, Supplementary Material online).
About 7%, 6%, 26%, and 5% of all genes in the genomes
of 2CP-C, 2CP-1, Fw109-5, and K, respectively, have no orthologs in any of the other Anaeromyxobacter genomes. Because the 2CP-C genome is the best-characterized of these
four genomes (Thomas et al. 2008) and appear to represent
well all four genomes, we used the A. dehalogenans 2CP-C
genome for our further analyses. To increase the resolving
power of comparative genome analyses, we determined
the complete nucleotide sequence of the S. aurantiaca
DW4-3-1 genome (table 1).
These five species cover all three Myxococcales suborders: M. xanthus, S. aurantiaca, and A. dehalogenans belong
to the Cystobacterineae, S. cellulosum belong to the Sorangineae, and H. ochraceum belong to the Nannocystineae
(Fig. 1 and table 1) (Shimkets et al. 2006). In all three suborders, species have been isolated, which initiate a developmental program resulting in fruiting body formation
(Shimkets et al. 2006).
Anaeromyxobacter dehalogenans is significantly different
from the four other species in several ways (table 1). Although the others are strict aerobes, A. dehalogenans is
MBE
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
Table 2. COG Categorization of Genes of the Five Myxococcales Species Analyzed in This Studya.
Functional category
RNA processing and modification
Chromatin structure and dynamics
Energy production and conversion
Cell cycle control, cell division, chromosome partitioning
Amino acid transport and metabolism
Nucleotide transport and metabolism
Carbohydrate transport and metabolism
Coenzyme transport and metabolism
Lipid transport and metabolism
Translation and ribosome structure
Transcription
Replication, recombination, and DNA repair
Cell wall, membrane, envelope biogenesis
Cell motility
Post-translation modification, protein turnover, chaperones
Inorganic ion transport and metabolism
Secondary metabolite biosynthesis, transport, catabolism
Signal transduction
Intracellular traffic, secretion, and vesicular transport
Defense mechanisms
Cytoskeleton
Hypotheticalb
a
b
Code
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
T
U
V
Z
X
Myxococcus
xanthus
5(0.1)
5(0.1)
262(3.6)
41(0.6)
315(4.3)
94(1.3)
192(2.6)
156(2.1)
204(2.8)
225(3.1)
340(4.6)
290(4.0)
340(4.6)
57(0.8)
273(3.7)
268(3.7)
241(3.3)
566(7.7)
94(1.3)
96(1.3)
1(<0.1)
3268(44.5)
Stigmatella
aurantiaca
4(<0.1)
3(<0.1)
257(3.1)
49(0.6)
324(3.9)
102(1.2)
306(3.7)
168(2.0)
188(2.3)
228(2.7)
352(4.2)
326(3.9)
329(3.9)
58(0.7)
272(3.3)
254(3.0)
261(3.1)
674(8.1)
93(1.1)
89(1.1)
1(<0.1)
4014(48.1)
Anaeromyxobacter
dehalogenans
2(<0.1)
2(<0.1)
296(6.8)
38(0.9)
231(5.3)
70(1.6)
143(3.3)
120(2.8)
127(2.9)
198(4.6)
214(4.9)
169(3.9)
257(5.9)
59(1.4)
187(4.3)
178(4.1)
93(2.1)
339(7.8)
63(1.4)
70(1.6)
0(0)
1490(34.3)
Sorangium
cellulosum
3(<0.1)
2(<0.1)
364(3.9)
62(0.7)
365(3.9)
97(1.0)
343(3.7)
209(2.2)
198(2.1)
231(2.5)
475(5.1)
323(3.4)
335(3.6)
39(0.4)
313(3.3)
298(3.2)
288(3.1)
831(8.9)
69(0.7)
98(1.0)
0(0)
4441(47.3)
Haliangium
ochraceum
2(<0.1)
3(<0.1)
238(3.5)
45(0.7)
294(4.4)
72(1.1)
181(2.7)
139(2.1)
173(2.6)
186(2.8)
249(3.7)
237(3.5)
235(3.5)
24(0.4)
205(3.1)
180(2.6)
192(2.8)
556(8.3)
59(0.8)
78(1.2)
1(<0.1)
3370(50.2)
Numbers indicate total number of genes in each COG in that species, with numbers in parentheses indicating percent of all proteins in that species.
The category hypothetical combines COG categories ‘‘general prediction’’ and ‘‘conserved hypothetical’’ and proteins for which no functional assignment was made.
microaerophilic and its genome size is roughly half that of
the other species’. The most significant difference, for the
purposes of this study, is that A. dehalogenans has not been
demonstrated to initiate fruiting body formation and sporulation in response to starvation (Sanford et al. 2002;
Thomas et al. 2008). The morphology of the fruiting bodies
formed by the four fruiters varies significantly (table 1):
Myxococcus xanthus fruiting bodies are haystack shaped,
S. aurantiaca produces elaborate fruiting bodies consisting
of a stalk carrying several spore-filled sporangioles, and S.
cellulosum as well as H. ochraceum generate assemblies of
sporangioles, which are not carried by a stalk. Although the
vegetative cells in all four species are long and rod shaped,
spore morphology varies from spherical in M. xanthus
and H. ochraceum to short rods in S. aurantiaca and
S. cellulosum (table 1).
To compare the functional gene repertoires of the five
species, functional category assignments were created for
all genes in each of the five species based on COGs (Tatusov
et al. 1997). For this analysis, the assignments were adjusted
such that each gene was assigned only one functional category (Materials and Methods). By assigning one functional
category to each gene and maintaining consistency of assignments among predicted orthologs, we were able to analyze the distribution and conservation of functions across
species in a standardized manner. In agreement with previous observations for M. xanthus and S. cellulosum
(Schneiker et al. 2007), all five organisms have similar relative COG distributions with the exception of genes specifying proteins of unknown function (category X), the
percentage of which is significantly lower in A. dehalogenans than in the four other species (table 2).
Gene Conservation Within the Myxococcales
To be able to analyze the significance of the conservation of
genes directly involved in fruiting body and sporulation in
M. xanthus in the four other species or in other bacteria, we
first established a baseline for these comparisons by systematically identifying possible orthologs for each Myxococcales gene (from here on predicted orthologs are
referred to as orthologs). To this end, we used a reciprocal
best BlastP hit method (Overbeek et al. 1999; Fang et al.
2010) in which each protein from each species was tested
against the protein repertoires of each of the four other
species as well as against the protein repertoires of 973 bacterial species for which the complete, annotated genome
was available at the time of this study (Materials and Methods). From these analyses, Myxococcales proteins were
sorted into two categories: Myxococcales specific and
non-Myxococcales specific, based on whether or not an
ortholog is found in a non-Myxococcales genome. The
M. xanthus, S. aurantiaca, A. dehalogenans, S. cellulosum,
and H. ochraceum genomes contain 1,240, 1,501, 707,
1,731, and 1,165 pairs of paralogous genes, respectively. Paralogs represent a potential problem in the identification of
orthologs using a reciprocal best BlastP hit method and
may result in an underestimation of orthologs (Overbeek
et al. 1999; Fang et al. 2010). Using a less-stringent definition of orthologs (Materials and Methods), we found that
the false-negative discovery rate for all pairwise genome
comparisons varied between 2% and 4% (data not shown).
We conclude that our method results in a valid estimation
of orthologs.
In agreement with the phylogenetic analysis, the protein
repertoires of M. xanthus and S. aurantiaca appear more
1087
MBE
Huntley et al. · doi:10.1093/molbev/msq292
Table 3. Percent of Orthologs in Pairs of Myxococcales Species.
% of Orthologs Found In
Gene Repertoire
M. xanthus
S. aurantiaca
A. dehalogenans
S. cellulosum
H. ochraceum
Myxococcus
xanthus
100
56
52
28
34
Stigmatella
aurantiaca
63
100
50
28
35
similar to each other than any other species pairing, sharing
well over half of their protein repertoires (table 3). In contrast, less than half of either M. xanthus or S. aurantiaca
proteins are conserved in A. dehalogenans. Given the small
size of the A. dehalogenans genome, this was not surprising.
In agreement with earlier comparisons of M. xanthus
and S. cellulosum (Schneiker et al. 2007), only about 30%
of M. xanthus and S. aurantiaca proteins are conserved
in S. cellulosum and in H. ochraceum and vice versa. Finally,
only approximately 30% of the protein repertoires of
S. cellulosum and H. ochraceum are shared.
Anaeromyxobacter
dehalogenans
31
26
100
19
23
Sorangium
cellulosum
36
31
41
100
35
Haliangium
ochraceum
31
28
36
25
100
Pairwise comparisons of the position of orthologous
genes on their respective genomes revealed a high degree
of conserved synteny between M. xanthus and S. aurantiaca genomes and a lesser degree of synteny between these
two genomes and that of A. dehalogenans (Fig. 2). The evident X-shaped pattern in the M. xanthus versus S. aurantiaca plot—which is also evident at a lower resolution in
the plots comparing M. xanthus or S. aurantiaca with
A. dehalogenans—likely reflects large chromosomal inversions that reverse gene order symmetrically around the origin of replication (Eisen et al. 2000). In all other pairwise
FIG. 2. Synteny plot for orthologous genes in pairwise combinations of genomes of Myxococcales species. Each dot indicates the position of an
orthologous gene pair on the indicated chromosomes. Chromosome orientations and sizes in megabases (Mb) are indicated on the appropriate
axis for each species. The dnaA gene is defined as gene 1 in all five species.
1088
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
MBE
Table 4. Gene Conservation in Five Myxococcales Genomes.
Genomeb
Segmenta
IV
III
II
I
M.x.
S.a.
A d.
Gene repertoirec
S.c.
H.o.
Five Myxococcales
Genomes
1056j1046
273j262
197j186
425j403
40j39
469j376
286j226
94j88
400j357
70j70
1548j886
33j32
133j118
254j221
159j131
1896j809
30j30
39j39
19j19
103j100
89j79
212j162
290j249
2916j1368
74j73
196j179
122j110
1506j986
534j363
5506j2558
2290j1521
M.x. Development
Genes
15j15
3j3
2j2
3j2
0
7j6
3j3
0
9j6
0
36j24
1j1
0
4j4
1j0
11j7
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Downregulated during
M.x. Development
160j160
18j17
9j9
20j19
2j2
33j27
17j13
4j3
16j14
1j1
65j39
0
2j2
8j7
8j7
61j30
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Upregulated during
M.x. Development
40j38
17j16
5j4
26j24
2j2
32j23
24j20
6j6
20j17
2j2
117j62
0
7j7
12j11
13j11
87j37
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NOTE.—M.x., Myxococcus xanthus; S.a., Stigmatella aurantiaca; A.d., Anaeromyxobacter dehalogenans; S.c., Sorangium cellulosum; and H.o., Haliangium ochraceum.
a
The four segments labeled I, II, III, and IV highlight M. xanthus–specific conservation, M. xanthus/S. aurantiaca–specific conservation, fruiter-specific conservation, and
conservation in all five Myxococcales species, respectively.
b
Gray shading indicates the presence of a gene in the genome of that species.
c
The left number indicates the number of genes shared between the species marked with grey shading for four different gene repertoires. The right number indicates the
subset of those proteins for which an ortholog was found in at least one non-Myxococcales species.
comparisons, no discernable synteny was observed
(Fig. 2). Thus, in agreement with the phylogenetic relatedness of the five Myxococcales, gene organization is significantly more conserved between M. xanthus, S. aurantiaca,
and A. dehalogenans compared with S. cellulosum
and H. ochraceum. Moreover, gene organization in S. cellulosum and H. ochraceum does not show a high degree of
conservation.
To analyze the divergence in the gene repertoires of the
five Myxococcales since their last common ancestor, we
identified genes conserved in all five species and genes having lineage specificity within the order. About 1,056 genes
are conserved between all five Myxococcales genomes, and
most of these genes (1,046; 99%) have orthologs outside the
Myxococcales (Segment IV in table 4 and supplementary
table S2, Supplementary Material online). The 1,056 genes
correspond to the Myxococcales core genome. The core
genome was found to represent from a high of 24% of
all genes in a species (A. dehalogenans) to a low of 11%
(S. cellulosum). In agreement with the idea that the core
genome is expected to represent functions required for
these bacteria to thrive despite differences in their respective ecological niches, 85% of the core genes have inferred
functions with the largest functional categories including
housekeeping proteins involved in translation and ribosome structure (J); amino acid transport and metabolism
(E); energy production and conversion (C); replication, recombination, and DNA repair (L); and cell wall, membrane,
and envelope biogenesis (M). Interestingly, the largest COG
category in the core genome is that of proteins of unknown
function (X) (15%). About 53 genes (5%) for signal transduction (T) are also part of the core genome. The 10 genes
in segment IV without orthologs outside the Myxococcales
include six for proteins of unknown function, and the rest
are involved in general cellular processes such as energy
production, metabolism, and transcription. The Myxococcales core genome concurs reasonably well with other core
genomes previously defined at different taxonomic levels.
For example, in the species Streptococcus agalactiae, the
core genome is estimated to consist of 1,806 genes, corresponding to 80% of any genome, and with 67% of the core
genes having an inferred function (Tettelin et al. 2005). In
1089
MBE
Huntley et al. · doi:10.1093/molbev/msq292
Table 5. Summary of Orthologs in Pairs of Myxococcales Species.
% Found In
Myxococcus xanthus Repertoire
All coding sequences
Development-specific genes
Genes downregulated during development
Genes upregulated during development
Myxococcus
xanthus
100
100
100
100
the genus Prochlorococcus, the core genome includes approximately 1,250 genes corresponding to 40–67% of
the genes of the 12 fully sequenced Prochlorococcus genomes, and 81% of the genes have an inferred function
(Kettler et al. 2007). Finally, in the order Lactobacillales,
the core genome is made up of 567 genes corresponding
to 20–33% of the genes in fully sequenced Lactobacillales
genomes, and 84% of these genes have an inferred function
(Makarova et al. 2006).
About 425 genes are conserved in M. xanthus, S.
aurantiaca, S. cellulosum, and H. ochraceum (Segment III
in table 4 and supplementary table S2, Supplementary Material online). These genes could, in principle, represent the
signature genes encoding the functions needed to be
a member of the fruiting body-forming Myxococcales.
Among the 425 genes, those involved in ion transport
and metabolism (P) (8%) as well as proteins of unknown
function (X) (30%) represent the two largest groups. In addition, these putative fruiting body–specific genes are enriched for signal transduction (T), amino acid transport and
metabolism (E), as well as energy production and conversion (C), functional categories. Interestingly, most of the
425 genes (403; 95%) have orthologs outside the Myxococcales. The 22 Myxococcales-specific genes are enriched for
those involved in signal transduction (T) as well as proteins
of unknown function (X).
Between the five Myxococcales species, 1,548 genes are
shared specifically by M. xanthus and S. aurantiaca (Segment II in table 4 and supplementary table S2, Supplementary Material online), and 57% of this group (886) have
orthologs outside the Myxococcales. Regardless of the presence or absence of non-Myxococcales orthologs, genes involved in signal transduction (T) as well as those of
unknown function (X) are enriched in this group. Finally,
26%, 35%, 35%, 59%, and 34% of all genes from M. xanthus,
S. aurantiaca, A. dehalogenans, S. cellulosum, and H. ochraceum, respectively, have no ortholog in any of the other
Myxococcales. Approximately, 50% of these genes have
an ortholog in a non-Myxococcales species. Like the genes
shared between M. xanthus and S. aurantiaca, M. xanthus–
specific genes with and without orthologs in non-Myxococcales species (Segment I in table 4 and supplementary
table S2, Supplementary Material online) are enriched for
those involved in signal transduction (T) as well as proteins
of unknown function (X).
Several patterns emerge from these analyses. First,
M. xanthus genes at the four different levels of conservation
(Segments I–IV in table 4) are differentially enriched for
1090
Stigmatella
aurantiaca
63
82
80
69
Anaeromyxobacter
dehalogenans
31
29
53
26
Sorangium
cellulosum
36
36
54
30
Haliangium
ochraceum
31
26
52
28
genes specifying different functional categories. Those conserved in all five species or by the four fruiters are dominated by genes for general cellular processes, whereas those
conserved in M. xanthus and M. xanthus plus S. aurantiaca
are dominated by signal transduction proteins and proteins
of unknown function. Second, there is a gradient of conservation with respect to orthologs outside the Myxococcales with most genes (99%) shared by all five species
having orthologs in non-Myxococcales but only 43% of
the genes specific to M. xanthus having orthologs in
non-Myxococcales.
Conservation of M. xanthus Development Genes
The previous analyses established a baseline from which to
analyze the extent of conservation of M. xanthus genes directly involved in fruiting body and sporulation in the three
other species. We identified 95 genes in M. xanthus that are
directly involved in fruiting body formation and/or sporulation but do not appear to influence cells growing vegetatively (from here on this gene set is referred to as
M. xanthus development genes) (supplementary table
S3, Supplementary Material online). It should be emphasized that 60 (64%) of these genes belong to the signal
transduction (T) category (supplementary table S3, Supplementary Material online), whereas only 7.7% of the total M.
xanthus proteins belong to this COG. Thus, the M. xanthus
development genes are subject to a sampling bias. Nevertheless, as a first approximation, we used this set of genes to
analyze the possible conservation of the M. xanthus genetic
program for fruiting body formation and sporulation in
other Myxococcales.
About 78 (82%) of the 95 M. xanthus development genes
are conserved in S. aurantiaca, 28 (29%) in A. dehalogenans,
34 (36%) in S. cellulosum, and 25 (26%) in H. ochraceum
(table 4). The M. xanthus development genes shared with
S. aurantiaca are overrepresented, compared with the overall frequency of M. xanthus genes having orthologs in
S. aurantiaca (82% vs. 63%) (tables 3 and 5 for summary).
The frequency of M. xanthus development genes shared
with A. dehalogenans (29% vs. 31%) or with S. cellulosum
(36% vs. 36%) equals the frequency of all M. xanthus genes
shared with either of these two organisms. Myxococcus xanthus development genes shared with H. ochraceum (31% vs.
26%) are slightly underrepresented. With the caveat
that the M. xanthus development-specific genes are
subject to a sampling bias, these comparisons suggest that
M. xanthus development genes are overrepresented in
S. aurantiaca and not more likely to be conserved than
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
any other gene or slightly underrepresented in S. cellulosum
and H. ochraceum, respectively.
Interestingly, only 22 of the 95 M. xanthus developmentspecific genes are specific to the Myxococcales with the
remaining 77 having orthologs in non-Myxococcales species (table 4 and supplementary table S3, Supplementary
Material online). About 15 development-specific genes
are conserved in all five Myxococcales and they all have
orthologs outside Myxococcales. Only three genes are
shared by the four fruiters, two have orthologs outside
the Myxococcales. Three (M. xanthus, S. aurantiaca,
H. ochraceum), zero (M. xanthus, S. cellulosum, H. ochraceum),
and nine genes (M. xanthus, S. aurantiaca, S. cellulosum)
are specifically shared by the three different combinations
of three fruiters. About 36 genes are conserved only in
M. xanthus and S. aurantiaca, and of these, 24 are found outside the Myxococcales. Finally, 11 are specific to M. xanthus,
and of these, seven are found in a non-Myxococcales.
By parsimony, the M. xanthus development-specific
genes shared by all five Myxococcales likely existed in their
common ancestor. These 15 genes are all conserved outside
Myxococcales, suggesting that they could be part of an ancestral response to starvation rather than being specifically
involved in fruiting body formation and/or sporulation. In
agreement with this notion, the RelA protein is found in
this group. Notably, RelA is also important for fruiting body
formation in S. cellulosum (Knauber et al. 2008). RelA is required for induction of the stringent response in M. xanthus and in S. cellulosum and is a universally conserved
protein in bacteria. The stringent response is a ubiquitously
conserved mechanism in bacteria that links nutrient availability to accumulation of the second messenger molecular
guanosine-5#-(tri)di-3#-diphosphate ((p)ppGpp) by means
of the RelA and/or SpoT proteins (Potrykus and Cashel
2008). Similarly, genes shared by the four fruiters could
be relics of an ancient mechanism for fruiting body formation and sporulation and which was lost in A. dehalogenans.
In contrast, genes specific to M. xanthus and S. aurantiaca
may be part of a development mechanism specific to those
species’ lineage and might reflect an adaptation to their
particular lifestyle or environment.
The above interpretation is based on numbers. We
thought that a more biological approach to determine
the relevance of the differential conservation of M. xanthus
development-specific genes would be to determine to
which extent genes for specific signal transduction pathways
important for development in M. xanthus are conserved in
the four other species. Therefore, we analyzed the conservation of the genes in two of the best-studied developmental
pathways in M. xanthus, the MrpC/FruA pathway, and the
C-signal transduction pathway (Fig. 3). Briefly, in the MrpC/
FruA pathway, expression of the mrpAB genes encoding
a two-component system is induced by starvation (Sun
and Shi 2001a, 2001b). Phosphorylated MrpB induces expression of mrpC that encodes a transcriptional regulator
of the CRP family (Sun and Shi 2001a, 2001b). In vegetative
cells, MrpC is phosphorylated by the Pkn8/Pkn14 cascade
both of which are Ser/Thr protein kinases (Nariya and In-
MBE
ouye 2005). During starvation, phosphorylation of MrpC
is thought to cease and instead MrpC is thought to be
cleaved by the LonD protease to MrpC2 (Ueki and Inouye
2003). The EspA histidine protein kinase by an unknown
mechanism inhibits accumulation of MrpC (Higgs et al.
2008). MrpC2 activates transcription of fruA (Ueki and Inouye 2003) encoding a DNA-binding response regulator
(Ogawa et al. 1996; Ellehauge et al. 1998) and mazF encoding a toxin protein involved in programmed cell death during development (Nariya and Inouye 2008). Intriguingly, all
the genes for the proteins in this pathway except Pkn8 are
conserved in S. aurantiaca (Fig. 3A), whereas these genes
are largely absent in A. dehalogenans, S. cellulosum, and H.
ochraceum.
The same picture emerges from an analysis of the conservation of proteins in the C-signal transduction pathway
(Fig. 3B): Transcription of the csgA gene is induced in
response to starvation in a RelA-dependent manner
(Crawford and Shimkets 2000). The full-length CsgA protein (p25) is exported and anchored in the outer membrane (Lobedanz and Søgaard-Andersen 2003). Here, p25
is cleaved by the PopC protease to generate p17, the actual
C-signal (Kim and Kaiser 1990; Lobedanz and SøgaardAndersen 2003; Rolbetzki et al. 2008). Transcription of csgA
is also regulated by the four ActABCD proteins (Gronewold
and Kaiser 2001). C-signal transmission likely induces phosphorylation of FruA (Ellehauge et al. 1998). Phosphorylated
FruA inhibits the activity of the Frz chemosensory system
(Søgaard-Andersen and Kaiser 1996). Subsequently, the activity of the Ras-like GTPase MglA decreases and cells stop
reversing (Jelsbak and Søgaard-Andersen 2002) and eventually aggregate into nascent fruiting bodies. FruA in combination with LadA (Viswanathan et al. 2007b) stimulates
transcription of the devTRS operon (Viswanathan et al.
2007b), which encodes three clustered, regularly interspaced short palindromic repeat proteins-associated proteins (Viswanathan et al. 2007a), and alone induces
transcription of fdgA (Ueki and Inouye 2005). Several other
pathways converge on the C-signal transduction pathway
at the level of FruA phosphorylation. These pathways are
defined by proteins of two-components systems including
SdeK (Pollack and Singer 2001), TodK (Rasmussen and
Sogaard-Andersen 2003), RodK (Rasmussen et al. 2005),
and RokA (Wegener-Feldbrugge and Sogaard-Andersen
2009) as well as the transcription factor MXAN4899
(Jelsbak et al. 2005). Intriguingly, with the exception of
the genes for DevTRS, all these genes are conserved in
S. aurantiaca. On the other hand, only a few of the genes
in the C-signal transduction pathway are present in
A. dehalogenans, S. cellulosum, and H. ochraceum.
The conservation of essentially all the genes for the proteins
in these two pathways in S. aurantiaca suggests that these
pathways may also exist in S. aurantiaca and could be important for development in this organism. On the other hand, the
almost complete absence of the genes for the proteins of these
two pathways in S. cellulosum and H. ochraceum strongly suggests that these two pathways do not exist to regulate development in these two organisms.
1091
MBE
Huntley et al. · doi:10.1093/molbev/msq292
A
MrpB -P
FruA
MazF
fruA
mazF
Colour code:
M.x. S.a. A.d. S.c. H.o.
MrpC2
MrpA -P
MrpC
LonD
Signal?
-P
Pkn14 -P
MrpB
Pkn14
MrpC
MrpA
Pkn8
EspA
mrpA
mrpB
-P
Pkn8
mrpC
Signal?
Starvation
B
FrzF
FrzA
FrzCD
FrzE
MglA
FrzZ
Reversal
frequency
Aggregation
FrzB
FrzG
FruA
-P
LadA
p17
DevT
DevR
FdgA
DevS
Sporulation
PopC
ActA
ActB
ActC
ActD
actA
actB
actC
actD
devT
devR
devS
fdgA
p25
csgA
FruA
fruA
4899
SdeK
TodK
sdeK
todK
RokA
RodK
(p)ppGpp
RelA
Starvation
FIG. 3. Two signal transduction pathways important for development in Myxococcus xanthus. (A) The MrpC/FruA pathway. See the text for
details. Genes and proteins are color coded according to conservation as shown in the box. Proteins not circled by a black line are also
important for vegetative growth and are, therefore, not included in the list of M. xanthus–specific developmental genes. (B) The C-signal
transduction pathway. See the text for details. The color code is as in (A). The stippled line indicates the cell envelopes of two neighboring cells.
The gray box around FruA;P indicate that the mechanisms by which the MXAN_4899, SdeK, TodK, and RodK/RokA pathways converge with
the C-signal transduction pathway are unknown.
Because several of the M. xanthus development-specific
genes present only in M. xanthus and S. aurantiaca are also
conserved in non-Myxococcales species (supplementary
table S3, Supplementary Material online), we sought to
determine whether these genes were lost from A. dehalogenans, S. cellulosum, and H. ochraceum or transferred by
horizontal gene transfer (HGT) to a common ancestor of
M. xanthus and S. aurantiaca. Therefore, we analyzed the
phylogenetic distribution of orthologs and homologs of
each M. xanthus development-specific gene. As a criterion
for HGT, we used disagreements between the phylogenetic
trees of a gene and that of the species (Beiko et al. 2005).
As shown in supplementary figure S2 (Supplementary
1092
Material online), most M. xanthus development-specific
genes have orthologs or homologs in the four other
Myxococcales as well as in non-Myxococcales making
conclusions about HGT difficult. Only in the case of
hthA (MXAN_6889) does the distribution of orthologs
and homologs support that this gene was acquired
by HGT by a common ancestor of M. xanthus, S. aurantiaca, and A. dehalogenans. mspA (MXAN_2269), nfsH
(MXAN_3378), and fruE (MXAN_4486) represent
genes which are only conserved in M. xanthus and S.
aurantiaca. Interestingly, most of these genes are specifically important for sporulation in M. xanthus and not
for aggregation.
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
Conservation of Developmentally Regulated Genes
in M. xanthus
Because there is a sampling bias in our set of M. xanthus
development-specific genes, we sought an alternative
method to evaluate the extent of conservation of genes
important for M. xanthus development in other species.
Many genes with a function in M. xanthus development
are transcriptionally regulated during fruiting body formation (e.g., see Fig. 3). Therefore, to obtain an unbiased set of
genes potentially required for development in M. xanthus,
we identified at a global scale genes that are transcriptionally regulated in response to starvation, using data from
a previously published M. xanthus DNA microarray experiment (Shi et al. 2008). In this experiment, total RNA was
isolated from M. xanthus cells after 0, 2, 4, 6, 9, 12, 15, 18,
and 24 h of starvation on a solid surface. Using as a criterion
for developmental regulation, a 1.5-fold up- or downregulation in at least one time point during development, we
found that, over the course of development, 424 genes
were downregulated and 410 upregulated (supplementary
tables S5 and S6, Supplementary Material online) and that
the COG categories represented by up- and downregulated
genes are significantly different (supplementary table S7
and S8, Supplementary Material online).
In total, 338 (80%) of the 424 downregulated genes are
conserved in S. aurantiaca, 225 (53%) in A. dehalogenans,
229 (54%) in S. cellulosum, and 220 (52%) H. ochraceum
(table 4). Compared with the overall percentages of M. xanthus genes having orthologs in the three other Myxococcales (table 3; summary table 5), downregulated genes are
overrepresented in all species, suggesting the existence of
a common transcriptional response to starvation in these
five species. Moreover, a total of 350 (83%) of all downregulated genes have orthologs outside the Myxococcales
(table 4). Nearly half (160) of the downregulated genes
are conserved in all five Myxococcales and also conserved
outside the order. Three of the largest groups of downregulated genes shared by all five Myxococcales include the
categories translation and ribosome structure (J), energy
production and conversion (C), and posttranslation modification, protein turnover, chaperones (O), and, thus, likely
represent the stringent response in M. xanthus (supplementary table S7, Supplementary Material online). Moreover, a significant fraction of these genes represent
proteins of unknown function. There are relatively few
(20) downregulated genes conserved in the four fruiting
body forming bacteria and, of those, most are involved
in posttranslation modification, protein turnover, chaperones (O). Conversely, of the M. xanthus–specific and M.
xanthus and S. aurantiaca–specific genes downregulated
during development, more than 50% represent proteins
of unknown function (X).
In all, 280 (68%) of the 410 M. xanthus genes upregulated
during development have orthologs outside the Myxococcales; 281 (69%) of the 410 upregulated genes are found in
S. aurantiaca, 105 (26%) in A. dehalogenans, 125 (30%) in
S. cellulosum, and 116 (28%) in H. ochraceum (table 4).
MBE
Compared with the overall percentages of M. xanthus
genes having orthologs in the four other Myxococcales
(table 5 for summary), M. xanthus upregulated genes
shared with S. aurantiaca are overrepresented (69% vs.
63%). As expected, M. xanthus upregulated genes shared
with the non-developer A. dehalogenans are slightly underrepresented (26% vs. 31%). Strikingly, the percentage of M.
xanthus upregulated genes shared with S. cellulosum (30%
vs. 36%) or with H. ochraceum (28% vs. 31%) are also
slightly underrepresented.
In agreement with these observations, the conservation
of upregulated genes between the four Myxococcales is
very different compared with that of downregulated genes:
Half of the upregulated genes are found either only in M.
xanthus (87; 21%) or are shared with S. aurantiaca (117;
29%) (table 4). Most of these genes are classified as signal
transduction proteins or proteins of unknown function
(supplementary table S8, Supplementary Material online).
Only 26 (6%) of the upregulated genes are conserved in all
four fruiters and 40 (10%) genes are conserved in all five
species (table 4). Nearly all these more broadly conserved
genes within the Myxococcales are also found outside
the order (table 4), suggesting that these may be genes involved in starvation-induced stress response mechanisms
common to many bacteria.
Discussion
The aim of the work presented here was to investigate the
hypothesis that a common a genetic program underlies
fruiting body formation in Myxococcales. We used three
comparative genomics approaches to address this hypothesis. All approaches strongly indicate that the genetic programs for fruiting body formation in M. xanthus and
S. aurantiaca may share many components and perhaps
even entire signal transduction pathways in common.
All approaches also suggest that these programs in S. cellulosum and H. ochraceum are significantly different from
that in M. xanthus. Thus, the overall conclusion from these
analyses is that fruiting Myxococcales do not share
a common genetic program directing fruiting body formation. This conclusion is based on the following observations: First, a set of 95 M. xanthus development-specific
genes—although suffering from a sampling bias—are overrepresented in S. aurantiaca, whereas they occur at the
same frequency as an average M. xanthus gene in A. dehalogenans and in S. cellulosum and are underrepresented in
H. ochraceum. Second, essentially all proteins in the MrpC/
FruA and C-signal signal transduction pathways in M. xanthus are conserved in S. aurantiaca, whereas only a minority
of the proteins in these pathways are present in S. cellulosum, A. dehalogenans, or in H. ochraceum. Third, genes
upregulated during development in M. xanthus are overrepresented in S. aurantiaca and slightly underrepresented
in S. cellulosum, A. dehalogenans, and in H. ochraceum. In
contrast, genes downregulated during development in
M. xanthus are overrepresented in S. aurantiaca as well as
in S. cellulosum, A. dehalogenans, and H. ochraceum. These
1093
Huntley et al. · doi:10.1093/molbev/msq292
differences among up and downregulated genes suggest that
the five species studied could share in common a transcriptional response to starvation involving down-regulated
genes. The COG classification of the corresponding genes
and their high frequency of conservation outside the Myxococcales strongly suggest that this response represents the
stringent response.
The overall conclusion that fruiting Myxococcales do
not share a common genetic program directing fruiting
body formation has two interpretations. One would be that
the genetic programs governing fruiting body formation
evolved more than once in the Myxococcales. In other
words, the last common ancestor of the Myxococcales
was not a fruiting bacterium, fruiting bodies are not homologous structures, and fruiting body formation in the Myxococcales is an example of convergent evolution. All five
Myxococcales share M. xanthus development-specific proteins and upregulated genes in common. If fruiting body
formation did in fact evolve more than once, these shared
genes could be part of a general response to starvation. A
likely precedent for multicellularity evolving more than
once within an order are provided by the evolution of multicellularity in the green algae Volvocales, which form planar or spherical colonies. In this group of organisms, it has
been suggested that multicellularity has likely evolved
several times during evolution (Kirk 1999, 2005).
An alternative explanation for the overall conclusion that
fruiting Myxococcales do not share a common genetic
program directing fruiting body formation would be that
the last common ancestor of the Myxococcales was a fruiting
bacterium. Accordingly, this program would have been lost
in A. dehalogenans probably in parallel with a reduction in
genome size. A precedent for an apparently similar loss is
provided by the loss of the sporulation program in the order
Lactobacillales after the divergence from the last common
ancestor of Bacilli (Makarova et al. 2006). In M. xanthus
and S. aurantiaca, the extant version of this program would
largely be shared because their separation into separate
species occurred relatively late. In particular, the overrepresentation of M. xanthus development-specific proteins
and genes upregulated during development genes in
S. aurantiaca and the possible conservation of entire signal
transduction pathways in S. aurantiaca support this evolutionary scheme. Moreover, the overrepresentation of
M. xanthus development-specific genes and upregulated
genes in S. aurantiaca suggest that selection acts to maintain
these genes in M. xanthus and S. aurantiaca. In addition,
M. xanthus may have evolved development-specific functions as suggested by the existence of development-specific
proteins, which are not conserved in S. aurantiaca. We speculate that the same is true for S. aurantiaca. This idea is supported by the observation that fruiting body formation in
S. aurantiaca depends on the intercellular signaling molecule
stigmolone whereas this molecule has no effect on
M. xanthus development (Plaga et al. 1998). On the other
hand, the S. cellulosum and the H. ochraceum lineages separated much earlier from the M. xanthus and S. aurantiaca
lineage and during the course of evolution the initially shared
1094
MBE
program may have diverged to such an extent that the
extant overlap with the programs in M. xanthus and
S. aurantiaca is minor. According to this scheme, fruiting
bodies are homologous structures generated from nonhomologous proteins. On the basis of the existing data, we
cannot distinguish between the two scenarios, that is, convergent evolution or divergent evolution. Likewise, our data
do not allow conclusions about the extent to which the genetic programs for fruiting body formation in S. cellulosum
and H. ochraceum may overlap. Regardless, our findings
suggest considerable plasticity in and innovation of
developmental pathways in the Myxococcales.
The last common ancestor of the Myxococcales is
thought to have existed 2.0 Gyr ago (Hedges et al.
2006). The lack of conservation of a genetic program for
fruiting body formation and sporulation in Myxococcales
is in stark contrast to the conservation of the genetic programs for sporulation in bacilli and clostridia (both of
which are low GþC Gram-positive bacteria) and cell cycle
control in Alphaproteobacteria. Sporulation in bacilli,
which are generally facultative anaerobes, and clostridia,
which are obligate anaerobes, has been mostly studied
in B. subtilis. Except for the components of the phosphorelay that regulates initiation of sporulation in B. subtilis,
most other regulatory proteins in this program are conserved in sporulating Bacilli and Clostridia (Paredes et al.
2005). Clostridia are thought to have diverged from other
low G þ C Gram-positive bacteria 2.7 Gyr ago and bacilli 2.3
Gyr ago (Paredes et al. 2005), suggesting that the genetic
program for sporulation in these bacteria has been largely
conserved for 2.7 Gyr. Cell cycle control in Alphaproteobacteria has been mostly studied in Caulobacter crescentus.
A recent study of 13 proteins involved in cell cycle control
in C. crescentus showed that many of these are conserved in
other Alphaproteobacteria (Brilli et al. 2010), which had
their last common ancestor 2.0 Gyr ago (Hedges et al.
2006), suggesting that the genetic program for cell cycle
control in the Alphaproteobacteria was largely in place
2.0 Gyr ago. Cell cycle control is an essential process,
and therefore, the possibility of significant evolutionary
changes in this program is minor. On the other hand, sporulation is not absolutely essential for survival, thus possibly
allowing for a larger degree of change. However, this does
not explain the differences in level of conservation of proteins required for sporulation in the Myxococcales and low
G þ C Gram-positive bacteria.
The evolution of morphological variations in metazoans
has been suggested to have mainly involved cis-regulatory
mutations affecting gene expression patterns rather than
protein function (Stern and Orgogozo 2008; Tirosh et al.
2009) and gene duplication followed by neo- or subfunctionalization (Lynch and Conery 2000). Nevertheless,
there are examples in which the formation of homologous
structures are regulated by nonhomologous pathways
(De Robertis 2008; Sommer 2009). Likewise, orphan genes
have been suggested to have important functions in the
generation of morphological diversity (Khalturin et al.
2009). Bacterial genomes are thought to evolve mainly
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
by incorporation of novel genes by HGT, by intragenomic
duplications and by gene loss (Ochman et al. 2000; Lerat
et al. 2005; Dagan et al. 2008). It has previously been suggested that several genes important for development in M.
xanthus were acquired by HGT and ‘‘fashioned into
a unique developmental cycle and presumably played a different role in their previous host’’ (Goldman et al. 2007).
Our analyses do not support HGT as a significant source
for genes in developmental pathways in M. xanthus. These
different conclusions are likely based on the much larger set
of bacterial genomes available for our analyses.
In summary, our analyses reveal an unexpected level of
plasticity in the genetic programs for fruiting body formation and sporulation in the Myxococcales. Future work will
be directed at elucidating whether this plasticity is the result of divergent evolution of a homologous program or
convergent evolution of similar structures.
Supplementary Material
Supplementary tables S1–S8 and Figures S1–S2 are available at Molecular Biology and Evolution online (http://
mbe.oxfordjournals.org/).
Acknowledgments
We thank Penelope Higgs and Frank Müller for helpful discussions. US National Science Foundation grant number
0236595 (to W.C.N.) and the Max Planck Society supported
this work.
References
Beiko RG, Harlow TJ, Ragan MA. 2005. Highways of gene sharing in
prokaryotes. Proc Natl Acad Sci U S A. 102:14332–14337.
Brilli M, Fondi M, Fani R, Mengoni A, Ferri L, Bazzicalupo M,
Biondi E. 2010. The diversity and evolution of cell cycle
regulation in alpha-proteobacteria: a comparative genomic
analysis. BMC Syst Biol. 4:52.
Chevenet F, Brun C, Banuls AL, Jacq B, Christen R. 2006. TreeDyn:
towards dynamic graphics and annotations for analyses of trees.
BMC Bioinformatics. 7:439.
Crawford EW, Shimkets LJ. 2000. The Myxococcus xanthus socE and
csgA genes are regulated by the stringent response. Mol
Microbiol. 37:788–799.
Dagan T, Artzy-Randrup Y, Martin W. 2008. Modular networks and
cumulative impact of lateral transfer in prokaryote genome
evolution. Proc Natl Acad Sci U S A. 105:10039–10044.
De Robertis EM. 2008. Evo-Devo: variations on ancestral themes. Cell
132:185–195.
Delcher AL, Harmon D, Kasif S, White O, Salzberg SL. 1999.
Improved microbial gene identification with GLIMMER. Nucleic
Acids Res. 27:4636–4641.
Dworkin M. 1972. The myxobacteria: new directions in studies of
prokaryotic development. Crit Rev Microbiol. 1:435–452.
Eisen JA, Heidelberg JF, White O, Salzberg SL. 2000. Evidence for
symmetric chromosomal inversions around the replication
origin in bacteria. Genome Biol. 1:RESEARCH0011.
Ellehauge E, Norregaard-Madsen M, Søgaard-Andersen L. 1998.
The FruA signal transduction protein provides a checkpoint for
the temporal co-ordination of intercellular signals in Myxococcus
xanthus development. Mol Microbiol. 30:807–817.
MBE
Fang G, Bhardwaj N, Robilotto R, Gerstein MB. 2010. Getting started
in gene orthology and functional analysis. PLoS Comput Biol.
6:e1000703.
Flärdh K, Buttner MJ. 2009. Streptomyces morphogenetics: dissecting
differentiation in a filamentous bacterium. Nat Rev Microbiol.
7:36–49.
Fudou R, Jojima Y, Iizuka T, Yamanaka S. 2002. Haliangium
ochraceum gen. nov., sp. nov. and Haliangium tepidum sp.
nov.: novel moderately halophilic myxobacteria isolated from
coastal saline environments. J Gen Appl Microbiol. 48:109–115.
Goldman B, Bhat S, Shimkets LJ. 2007. Genome evolution and the
emergence of fruiting body development in Myxococcus
xanthus. PLoS One. 2:e1329.
Goldman BS, Nierman WC, Kaiser D, et al. (22 co-authors). 2006.
Evolution of sensory complexity recorded in a myxobacterial
genome. Proc Natl Acad Sci U S A. 103:15200–15205.
Gordon D. 2003. Viewing and editing assembled sequences using
Consed. Curr Protoc Bioinformatics. 11:2.1–2.43.
Gronewold TM, Kaiser D. 2001. The act operon controls the level
and time of C-signal production for Myxococcus xanthus
development. Mol Microbiol. 40:744–756.
Hedges SB, Dudley J, Kumar S. 2006. TimeTree: a public knowledgebase of divergence times among organisms. Bioinformatics
22:2971–2972.
Higgs PI, Jagadeesan S, Mann P, Zusman DR. 2008. EspA, an orphan
hybrid histidine protein kinase, regulates the timing of
expression of key developmental proteins of Myxococcus
xanthus. J Bacteriol. 190:4416–4426.
Jelsbak L, Givskov M, Kaiser D. 2005. Enhancer-binding proteins with
a forkhead-associated domain and the sigma54 regulon in
Myxococcus xanthus fruiting body development. Proc Natl Acad
Sci U S A. 102:3010–3015.
Jelsbak L, Søgaard-Andersen L. 2002. Pattern formation by a cell
surface-associated morphogen in Myxococcus xanthus. Proc
Natl Acad Sci U S A. 99:2032–2037.
Kettler GC, Martiny AC, Huang K, et al. (14 co-authors). 2007.
Patterns and implications of gene gain and loss in the evolution
of Prochlorococcus. PLoS Genet. 3:e231.
Khalturin K, Hemmrich G, Fraune S, Augustin R, Bosch TCG. 2009.
More than just orphans: are taxonomically-restricted genes
important in evolution? Trends Genetics. 25:404–413.
Kim SK, Kaiser D. 1990. C-factor: a cell-cell signaling protein
required for fruiting body morphogenesis of M. xanthus. Cell
61:19–26.
Kirk DL. 1999. Evolution of multicellularity in the volvocine algae.
Curr Opin Plant Biol. 2:496–501.
Kirk DL. 2005. A twelve-step program for evolving multicellularity
and a division of labor. BioEssays 27:299–310.
Knauber T, Doss SD, Gerth K, Perlova O, Müller R, Treuner-Lange A.
2008. Mutation in the rel gene of Sorangium cellulosum affects
morphological and physiological differentiation. Mol Microbiol.
69:254–266.
Kroos L. 2007. The Bacillus and Myxococcus developmental
networks and their transcriptional regulators. Annu Rev Genet.
41:13–39.
Lagesen K, Hallin P, Rodland EA, Staerfeldt HH, Rognes T,
Ussery DW. 2007. RNAmmer: consistent and rapid annotation
of ribosomal RNA genes. Nucleic Acids Res. 35:3100–3108.
Leonardy S, Bulyha I, Søgaard-Andersen L. 2008. Reversing cells and
oscillating proteins. Mol BioSystems. 4:1009–1014.
Lerat E, Daubin V, Ochman H, Moran NA. 2005. Evolutionary origins
of genomic repertoires in bacteria. PLoS Biol. 3:e130.
Lobedanz S, Søgaard-Andersen L. 2003. Identification of the C-signal,
a contact dependent morphogen coordinating multiple developmental responses in Myxococcus xanthus. Genes Dev.
17:2151–2161.
1095
Huntley et al. · doi:10.1093/molbev/msq292
Lowe TM, Eddy SR. 1997. tRNAscan-SE: a program for improved
detection of transfer RNA genes in genomic sequence. Nucleic
Acids Res. 25:955–964.
Lynch M, Conery JS. 2000. The evolutionary fate and consequences
of duplicate genes. Science 290:1151–1155.
Makarova K, Slesarev A, Wolf Y, et al. (50 co-authors). 2006.
Comparative genomics of the lactic acid bacteria. Proc Natl Acad
Sci U S A. 103:15611–15616.
Myers EW, Sutton GG, Delcher AL, et al. (29 co-authors). 2000.
A whole-genome assembly of Drosophila. Science
287:2196–2204.
Nariya H, Inouye M. 2008. MazF, an mRNA interferase, mediates
programmed cell death during multicellular Myxococcus
development. Cell 132:55–66.
Nariya H, Inouye S. 2005. Identification of a protein Ser/Thr kinase
cascade that regulates essential transcriptional activators in
Myxococcus xanthus development. Mol Microbiol. 58:367–379.
Natalia I, Chris D, Elke L, et al. (33 co-authors). 2010. Complete
genome sequence of Haliangium ochraceum type strain (SMP2T). Stand Genomic Sci. 2:96–106.
O’Connor KA, Zusman DR. 1991a. Development in Myxococcus
xanthus involves differentiation into two cell types, peripheral
rods and spores. J Bacteriol. 173:3318–3333.
O’Connor KA, Zusman DR. 1991b. Behaviour of peripheral rods and
their role in the life cycle of Myxococcus xanthus. J Bacteriol.
173:3342–3355.
Ochman H, Lawrence JG, Groisman EA. 2000. Lateral gene transfer
and the nature of bacterial innovation. Nature 405:299–304.
Ogawa M, Fujitani S, Mao X, Inouye S, Komano T. 1996. FruA,
a putative transcription factor essential for the development of
Myxococcus xanthus. Mol Microbiol. 22:757–767.
Overbeek R, Fonstein M, D’Souza M, Pusch GD, Maltsev N. 1999.
The use of gene clusters to infer functional coupling. Proc Natl
Acad Sci U S A. 96:2896–2901.
Paredes CJ, Alsaker KV, Papoutsakis ET. 2005. A comparative
genomic view of clostridial sporulation and physiology. Nat Rev
Microbiol. 3:969–978.
Perez JC, Groisman EA. 2009. Evolution of transcriptional regulatory
circuits in bacteria. Cell 138:233–244.
Plaga W, Stamm I, Schairer HU. 1998. Intercellular signaling in
Stigmatella aurantiaca: purification and characterization of
stigmolone, a myxobacterial pheromone. Proc Natl Acad Sci
U S A. 95:11263–11267.
Pollack JS, Singer M. 2001. SdeK, a histidine kinase required for
Myxococcus xanthus development. J Bacteriol. 183:3589–3596.
Potrykus K, Cashel M. 2008. (p)ppGpp: still magical? Annu Rev
Microbiol. 62:35–51.
Rabus R, Kube M, Beck A, Widdel F, Reinhardt R. 2002. Genes
involved in the anaerobic degradation of ethylbenzene in
a denitrifying bacterium, strain EbN1. Arch Microbiol.
178:506–516.
Rasmussen AA, Porter SL, Armitage JP, Sogaard-Andersen L. 2005.
Coupling of multicellular morphogenesis and cellular differentiation by an unusual hybrid histidine protein kinase in.
Myxococcus xanthus. Mol Microbiol. 56:1358–1372.
Rasmussen AA, Sogaard-Andersen L. 2003. TodK, a putative
histidine protein kinase, regulates timing of fruiting body
morphogenesis in Myxococcus xanthus. J Bacteriol.
185:5452–5464.
Rolbetzki A, Ammon M, Jakovljevic V, Konovalova A, SogaardAndersen L. 2008. Regulated secretion of a protease activates
intercellular signaling during fruiting body formation in
M. xanthus. Dev Cell. 15:627–634.
Rosenbluh A, Nir R, Sahar E, Rosenberg E. 1989. Cell-densitydependent lysis and sporulation of Myxococcus xanthus in
agarose beads. J Bacteriol. 171:4923–4929.
1096
MBE
Rutherford K, Parkhill J, Crook J, Horsnell T, Rice P, Rajandream MA,
Barrell B. 2000. Artemis: sequence visualization and annotation.
Bioinformatics 16:944–945.
Sanford RA, Cole JR, Tiedje JM. 2002. Characterization and
description of Anaeromyxobacter dehalogenans gen. nov., sp.
nov., an aryl-halorespiring facultative anaerobic Myxobacterium.
Appl Environ Microbiol. 68:893–900.
Schneiker S, Perlova O, Kaiser O, et al. (58 co-authors). 2007.
Complete genome sequence of the myxobacterium Sorangium
cellulosum. Nat Biotechnol. 25:1281–1289.
Shi X, Wegener-Feldbrugge S, Huntley S, Hamann N, Hedderich R,
Sogaard-Andersen L. 2008. Bioinformatics and experimental
analysis of proteins of two-component systems in Myxococcus
xanthus. J Bacteriol. 190:613–624.
Shimkets L, Dworkin M, Reichenbach H. 2006. The myxobacteria. In:
Dworkin M, Falkow S, Rosenberg E, Schleifer K-H, Stackenbrandt E,
editors. The prokaryotes. New York: Springer. p. 31–115.
Shimkets LJ, Brun YV. 2000. Prokaryotic development: strategies to
enhance survival. In: Brun YV, Shimkets LJ, editors. Prokaryotic
development. Washington DC: American Society for Microbiology.
p. 1–7.
Søgaard-Andersen L, Kaiser D. 1996. C factor, a cell-surfaceassociated intercellular signaling protein, stimulates the cytoplasmic Frz signal transduction system in Myxococcus xanthus.
Proc Natl Acad Sci U S A. 93:2675–2679.
Sommer RJ. 2009. The future of evo-devo: model systems and
evolutionary theory. Nat Rev Genet. 10:416–422.
Stern DL, Orgogozo V. 2008. The loci of evolution: how predictable
is genetic evolution? Evolution 62:2155–2177.
Sun H, Shi W. 2001a. Genetic studies of mrp, a locus essential for
cellular aggregation and sporulation of Myxococcus xanthus.
J Bacteriol. 183:4786–4795.
Sun H, Shi W. 2001b. Analyses of mrp genes during Myxococcus
xanthus development. J Bacteriol. 183:6733–6739.
Tatusov RL, Koonin EV, Lipman DJ. 1997. A genomic perspective on
protein families. Science. 278:631–637.
Tettelin H, Masignani V, Cieslewicz MJ, et al. (46 co-authors). 2005.
Genome analysis of multiple pathogenic isolates of Streptococcus agalactiae: implications for the microbial ‘‘pan-genome’’.
Proc Natl Acad Sci U S A. 102:13950–13955.
Thomas SH, Wagner RD, Arakaki AK, Skolnick J, Kirby JR,
Shimkets LJ, Sanford RA, Loffler FE. 2008. The mosaic genome
of Anaeromyxobacter dehalogenans strain 2CP-C suggests an
aerobic common ancestor to the delta-proteobacteria. PLoS
One. 3:e2103.
Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improving
the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties
and weight matrix choice. Nucleic Acids Res. 22:4673–4680.
Tirosh I, Reikhav S, Levy AA, Barkai N. 2009. A yeast hybrid provides
insight into the evolution of gene expression regulation. Science
324:659–662.
Ueki T, Inouye S. 2003. Identification of an activator protein
required for the induction of fruA, a gene essential for fruiting
body development in Myxococcus xanthus. Proc Natl Acad Sci
U S A. 100:8782–8787.
Ueki T, Inouye S. 2005. Identification of a gene involved in
polysaccharide export as a transcription target of FruA, an
essential factor for Myxococcus xanthus development. J Biol
Chem. 280:32279–32284.
Viswanathan P, Murphy K, Julien B, Garza AG, Kroos L. 2007a.
Regulation of dev, an operon that includes genes essential for
Myxococcus xanthus development and CRISPR-associated genes
and repeats. J Bacteriol. 189:3738–3750.
Viswanathan P, Ueki T, Inouye S, Kroos L. 2007b. Combinatorial
regulation of genes essential for Myxococcus xanthus
Evolution of fruiting Fruiting body Body formation Formation · doi:10.1093/molbev/msq292
development involves a response regulator and a LysR-type
regulator. Proc Natl Acad Sci U S A. 104:7969–7974.
Wegener-Feldbrugge S, Sogaard-Andersen L. 2009. The atypical
hybrid histidine protein kinase RodK in Myxococcus xanthus:
spatial proximity supersedes kinetic preference in phosphotransfer reactions. J Bacteriol. 191:1765–1776.
MBE
Wireman JW, Dworkin M. 1977. Developmentally induced autolysis
during fruiting body formation by Myxococcus xanthus. J
Bacteriol. 129:798–802.
Wu D, Hugenholtz P, Mavromatis K, et al. (34 co-authors). 2009. A
phylogeny-driven genomic encyclopaedia of bacteria and
archaea. Nature 462:1056–1060.
1097