MBE Advance Access originally published online on April 2, 2008
Molecular Biology and Evolution 2008 25(7):1274-1281; doi:10.1093/molbev/msn076
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Novel Transcriptome Patterns Accompany Evolutionary Restoration of Defective Social Development in the Bacterium Myxococcus xanthus
* Max-Planck-Institute for Developmental Biology, Germany
Max-Planck-Institute for Terrestrial Microbiology, Germany
Department of Biology, Indiana University
E-mail: gvelicer{at}indiana.edu.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
Evolutionary trait losses can be restored by direct reversion or by compensatory pathways. Upon starvation, the bacterium Myxococcus xanthus develops into spore-bearing fruiting bodies, but this ability can be rapidly lost during evolution. Some developmentally defective strains of M. xanthus "cheat" on proficient strains during development by superior sporulation in mixed cultures. Here, we examine transcriptomic patterns accompanying the evolution of a cheater (obligate cheater [OC]) to a developmentally competent strain (PX) by a single mutation. Using quantitative real-time–polymerase chain reaction analysis of 5 genes essential for development, we initially show that restoration of development in strain PX was associated with increased expression of 4 of these genes, not only relative to OC but also relative to the developmentally proficient ancestor of both OC and PX (wild type [WT]). Global transcriptome analyses showed further that developmental expression of well more than 100 genes differ significantly between PX and the proficient WT ancestor. Moreover, the expression profile of PX was found to differ from that of WT more than does that of the defective intermediate strain OC. These results show that the restoration of a complex trait is accompanied by novel expression patterns across a large number and wide variety of genes, rather than by a large-scale return to ancestral expression patterns.
Key Words: Myxococcus xanthus • social development • reverse evolution • transcriptome
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
Phenotypic traits can be adaptive in some environments but maladaptive in others. Evolutionary adaptation to a novel set of selective conditions may result in decreased fitness in an environment to which a population had previously adapted (Domingo and Holland 1997
; Crill et al. 2000). Most dramatically, traits beneficial in previous environments might be completely lost by negative pleiotropic effects of mutations that are adaptive under different conditions (Velicer et al. 1998; Cooper and Lenski 2000; Spiers et al. 2002; MacLean et al. 2004). Consider a population previously adapted to environment A, which then adapts to environment B by beneficial mutations that cause loss of a trait that is beneficial in A. If this population later reencounters and readapts to environment A (or a selectively similar environment), it may or may not be capable of reevolving the A-beneficial phenotype that had been lost in B. The likelihood of such phenotypic reversion would depend on the precise nature of the mutations causing the loss (e.g., deletion/insertion vs. base-pair substitution) and whether the basic phenotype can be regained by compensatory mutations rather than by direct reversion to the ancestral genotype.
Several evolutionary reversions of fitness levels or phenotypes have been documented (Teotonio and Rose 2001; Whiting et al. 2003), including instances of both reversion via compensatory mutations (Burch and Chao 1999; Velicer and Yu 2003; Heineman et al. 2005) and direct mutation back to an ancestral allele (Crill et al. 2000). In one case, strains of the social bacterium Myxococcus xanthus were made defective in social motility via deletion of the pilin gene (pilA) that is essential for the normal social motility. Derivative populations were then allowed to evolve under conditions favoring social motility. All 8 populations improved at least minimally in their swarm expansion rates but 2 lineages showed particularly dramatic restoration of social swarming functions. This was accomplished by a fundamentally novel genetic pathway that did not involve restoration of pilin production (Velicer and Yu 2003). Similarly, T7 bacteriophage was made partially defective at host cell lysis by deletion of the gene for lysozyme production (Heineman et al. 2005). Experimental populations of this defective phage reevolved the ability to lyse cells effectively, but did so primarily by novel mutations in a gene not previously known to affect lysis.
Alternatively, mutations beneficial in one environment but detrimental in another can revert directly upon switching environments (Levin et al. 2000). This occurs when bacteriophage X174 adapts to the novel host Salmonella enterica and thereby incurs adaptive mutations that are harmful for growth on its normal host, Escherichia coli, but which later revert directly back to their ancestral states when the phages are forced to readapt to E. coli (Crill et al. 2000). With the advent of efficient sequencing and mutation identification technologies, microbial evolution experiments allow studies of evolutionary reversions at both phenotypic and molecular levels. Here, we focus on a dramatic phenotypic reversion in M. xanthus in which defective multicellular development was restored by a single compensatory mutation. We show that this evolutionary restoration of a cooperative trait is associated with major shifts in the developmental expression of numerous genes relative to the ancestral state.
Myxococcus xanthus is a predatory, soil-dwelling social bacterium that undergoes multicellular development in response to amino acid deprivation. Using directed movement to gather at high-density aggregation loci, 
1,00,000 cells cooperatively build fruiting body structures (Kroos et al. 1986; Shimkets 1999). A minority of the aggregating population differentiates into stress-resistant spores under standard laboratory conditions, and most of the remaining cells appear to undergo autolysis. Upon return to conditions with sufficient nutrients, spores germinate to form vegetative cells. Myxococcus xanthus development is initiated upon detection of nutrition limitation at a high cell density. Nutrient limitation induces the stringent response and, thus, accumulation of the intracellular signaling molecule (p)ppGpp (Singer and Kaiser 1995; Harris et al. 1998). In addition, 2 stage-specific intercellular signals, the A- and C-signals, synchronize cells and induce the major changes in gene expression patterns required for fruiting body formation (fig. 1) (Kaiser 2004).
|
The genetically programmed process of M. xanthus social development can evolve rapidly. In an earlier study, cheating genotypes evolved from a developmentally proficient wild-type ancestor during 1,000 generations in a liquid environment where sociality was not required for competitive viability (Velicer et al. 2000
). One such cheater strain, here termed obligate cheater (OC), is defective at development in clonal culture but sporulates more efficiently than wild type (WT) when present as a minority in a mixed population with WT. In a subsequent competition experiment (Fiegna and Velicer 2003), OC was mixed (1%) with WT (99%) and allowed to compete through several sequential rounds of alternating growth and development. After the fourth round of development, a descendant of OC (here termed "PX" for Phoenix) reevolved independent developmental proficiency (fig. 2) (Fiegna et al. 2006). Thus, the evolutionary transition from OC to PX represents an "escape" from a state of obligate social dependence (in which OC requires the presence of WT to make spores) to restored social independence. Moreover, the regained social performance of strain PX is superior to its ancestral genotypes in both pure culture sporulation assays and in direct competitions in which PX is mixed pairwise with both ancestors (Fiegna et al. 2006).
|
Given the few generations (
60) it took PX to evolve from OC, it seemed likely that the transition was accomplished by a single mutation of large phenotypic effect. This hypothesis was tested by sequencing the entire genome of PX, finding all mutational discrepancies relative to the previously sequenced genome of WT, and identifying any discrepancies unique to PX that are not shared with its intermediate ancestor OC (Velicer et al. 2006). Out of 15 total discrepancies between WT and PX, only one is unique to PX, whereas 14 are shared with OC. This single mutation is compensatory and was shown to be the sole cause of the OC to PX phenotypic transition. This mutation (hereafter referred to as the PX mutation) appears to be regulatory, occurring 128 bp upstream from the start site of a predicted GCN5 N-acetyltransferase (GNAT) family acetyltransferase encoded by MXAN1079 and downstream from the predicted coding sequence of the upstream gene (MXAN1078). Compensatory mutations causing phenotypic reversions by altering gene regulation rather than protein sequence might restore relevant expression patterns to an ancestral state, thereby restoring the ancestral phenotype or function. Alternatively, such regulatory mutations might cause phenotypic reversion by generating a novel gene expression profile. Here, we have evaluated these contrasting hypotheses by using 2 experimental approaches. We first analyzed the expression of 5 genes essential for development using quantitative real-time–polymerase chain reaction (qRT-PCR) and then performed global transcriptional profiling with DNA microarrays to compare developmental RNA expression profiles across the 3 strains WT, OC, and PX. Due to the distinct developmental phenotypes exhibited by these 3 strains, we expected to find major differences in developmentally regulated gene expression between OC and the 2 developmentally proficient strains but we especially sought to compare the expression profiles of WT and PX. To address the question of how the transcriptome of an organism might shift during a compensatory adaptation, the microarray portion of this study focuses on comparison of overall gene expression patterns during development rather than on speculation about molecular mechanisms by which the PX transition was brought about.
Our analyses show that a large number of genes are differentially regulated between at least 2 of the 3 strains. Importantly, the developmental gene expression profile of PX not only differs from that of its WT ancestor at many genes but also actually deviates more from WT than does the intermediate cheater OC. Thus, the restoration of a major cooperative trait by a single compensatory mutation was accompanied by a novel transcriptome pattern rather than by a return to the ancestral state. This result highlights the ability of organisms to evolve complex and socially adaptive traits using different gene regulation profiles.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
Strains
The 3 strains compared in this study are termed here WT, OC, and PX (Phoenix) (fig. 2). WT (GJV10) is a variant of GJV1 that is developmentally proficient and is marked with kanamycin resistance gene due to chromosomal integration of the plasmid pDW79 (Wall et al. 1999
). (Note that strain GJV1 was referred to as WT in [Fiegna et al. 2006; Velicer et al. 2006]. We use WT for GJV10 here for simplicity.) GJV10 was used rather than GJV1 to control for any possible effects of pDW79, which is also integrated into the chromosome of strain OC. GVB207.3 evolved from GJV1 and was also marked by transformation with pDW79 to generate strain OC, which is a developmentally defective OC (strain S2/pDW79/S- in [Velicer et al. 2000]). PX is a socially independent genotype that evolved from OC (Fiegna et al. 2006).
Development of Cells for qRT-PCR and Microarray Experiments
Cells were grown in CTT liquid medium (Hodgkin and Kaiser 1977) at 32 °C to a density of 5 x 108 cells/ml. To initiate development, cells were harvested and resuspendend in TPM buffer (10 mM Tris–HCl [pH 7.6], 1 mM KH2PO4, 8 mM MgSO4) to a calculated density of 5 x 109 cells/ml. For each strain, 100-µl aliquots of concentrated cell suspension were deposited on separate TPM plates for each time point. Cells were harvested at several developmental time points (e.g. 6, 12, 18 and 24 h for the microarray experiments) into 1.5 ml TPM liquid and transferred immediately into 6 ml of RNA Protect Bacteria Reagent (Qiagen, Hilden, Germany). Cells were incubated at room temperature for 10 min, harvested by centrifugation at 4,000 x g for 10 min (4 °C), and frozen in liquid N2 after removal of the supernatant. For the T0 sample, a sample of the original liquid culture used to initiate development was directly resuspended in RNA Protect Bacteria Reagent and processed in the same manner.
RNA Isolation
To isolate RNA, frozen pellets were thawed and resuspended in 4 ml of 30 mg/ml lysozyme (Sigma, St. Louis, MO) prepared in Tris-EDTA (TE) buffer (10 mM Tris-HCl; 1 mM ethylenediaminetetraacetic acid [EDTA], pH 8.0) for cell lysis. RNA extraction was carried out using the RNeasy Midi Kit (Qiagen). To remove DNA, RNA eluted in 250 µl of RNase free water was supplemented with 25 µl DNase buffer and 12.5 µl of DNase I (Ambion, Darmstadt, Germany) and incubated at 37 °C for 1 h. RNA was repurified using RNeasy clean-up kit (Qiagen). The absence of DNA in the RNA preparations was verified by polymerase chain reaction (PCR) using the primers 5'-CTCACTCACCCATAAAATCACCC (forward) and 5'-CCTCGAAGGCCTCCTGGA (reverse) for MXAN3668 in the M. xanthus genome (http://www.tigr.org).
Quantitative Real-Time–Polymerase Chain Reaction
Two micrograms of DNA-free RNA was reverse transcribed using the cDNA Archive Kit and the supplied hexamers according to the protocol recommended by the supplier (Applied Biosystems, Foster City, CA). In the next step, SYBR Green PCR Master Mix was added to cDNA from the reverse transcription of 2 ng RNA together with 100 nM (final concentration) of each of the 2 primers in a primer pair (cf., below). The qRT-PCR was performed on a 7300 Real Time PCR System (Applied Biosystems) using standard conditions. All primer pairs were designed to give DNA fragment with sizes of 125–150 bp using PrimerExpress as recommended by the supplier (Applied Biosystems). The cDNA was prepared from RNA isolated from 3 independent biological experiments. All 3 strains were developed in parallel in each biological replicate. For the RNA isolated from each biological experiment, 2 or 3 technical replicates were performed. The average values of gene expression estimates for within-block technical replicates were used as statistically independent data points. Thus, gene expression estimates for each gene/time point combination were based on 3 independent biological replicate values that each represents the within-block average of technical replicates for each combination. The primer pairs used for each gene are as follows: (forward/reverse) sdeK, 5'-GGTCGTCTTGCGGAACGT and 5'-TGGCGGAGACACACTCGAA; csgA, 5'-GGCCGCGCTGAACATG and 5'-GAGCAGCACGGTGACGAA; fruA, 5'-ATCATCTCGCAGTGCTTCGA and 5'-CCTCGGACCAGGGAGTTGA; devR, 5'-AAACATCACCAGCCTCCAGAA and 5'-TGCATGGCTCCTGCTCATT; and MXAN3227, 5'-ATGAACCTCTATCCGGACATCGT and 5'-AGCTCGAAGGCCGTCTCA.
Microarray Assays and Experimental Design
RNA was isolated from each strain at 0, 6, 12, 18, and 24 h after the onset of starvation using the method described above. A common reference sample was prepared by pooling equal amounts of WT RNA from each of the 5 harvesting time points. This common reference sample was subsequently used in all hybridization experiments. To analyze gene expression during development, an experimental sample was prepared for each strain consisting of RNA pooled in equal amounts from the 4 developmental time points (6, 12, 18, and 24 h). A total of 20 µg of RNA of this sample from each strain were cohybridized with the same amount of the common reference sample. Each cohybridization was performed on RNA isolated from 3 independent biological replicates.
Microarray Processing
The M. xanthus DNA microarray covers 88% of M. xanthus genes, and each gene is represented by a PCR product of 275–325 bp. The PCR products were spotted on glass slides that had previously been coated in a poly-L-lysine solution (0.0086% poly-L-lysine [Sigma]; 0.1x phosphate-buffered saline). Microarrays were postprocessed as described previously (Jakobsen et al. 2004; Overgaard et al. 2006).
cDNA Synthesis, Fluorescent Labeling, and Hybridization
Synthesis of cDNA was carried out using 20 µg of DNA-free total RNA for each experimental sample and 20 µg of common reference sample as described with minor modifications (Jakobsen et al. 2004). Briefly, cDNA synthesis was carried out in the presence of 0.5 mM dATP, dCTP, dGTP; 0.1 mM dTTP; and 0.4 mM aminoallyl dUTP in a total volume of 30 µl. After cDNA synthesis, samples were hydrolyzed with 10 µl 0.5 M EDTA (pH 8) and 10 µl 1 M NaOH, incubated for 15 min at 65 °C followed by addition of 10 µl 1 M HCl. The cDNA was purified using a Zymo kit (Zymo Research, Freiburg, Germany), vacuum dried, and resuspended in 13 µl of fresh 100 mM sodium bicarbonate, pH 9. The reference probe was labeled with Cy5 and the experimental probe with Cy3 as described (Jakobsen et al. 2004). Hybridizations were carried out as described (Overgaard et al. 2006).
Data Acquisition and Analysis
Microarrays were scanned simultaneously at 2 wavelengths (Cy3, 532 nm and Cy5, 632 nm) using a GenePix 4000B microarray scanner (Axon Instruments, Sunnyvale, CA). Image analysis and data processing were performed using the Genepix 6.0 software package (Axon Instruments). The ratio-normalized data set (mean ratio of medians = 1) containing median signal intensity and median signal background from each channel was further analyzed using Acuity 4.0 software (Axon Instruments) and the Significance Analysis of Microarrays software version 2.23 (SAM v.2.23), which assigns a score to each feature on a microarray on the basis of changes in gene expression relative to the standard deviation of repeated measurements (Tusher et al. 2001). A filtered subset of all features printed on the array was selected for analysis based on the following criteria: 1) they were identified by the Genepix 6.0 spot-finding algorithm ("Flags" >= 0) and 2) local background subtracted median intensity values of either the Cy3 (532 nm) or the Cy5 (635 nm) channel greater than 500. Only features with data points in all slides (i.e., for all 3 strains in all 3 experimental blocks) were included in the further analyses. This selection process left 6021 of 6565 genes for analysis, representing 81% of all protein-coding genes in the M. xanthus genome.
For statistical significance analysis of the filtered data points, we used SAM v2.23 to calculate a t-like statistic based on the variance of the data across the 3 experimental blocks. The cutoff delta value of the SAM analysis was chosen to correspond with a median false discovery rate of 10% for calls of significant expression differences between strains. Using this cutoff criterion, SAM identified 188 genes that were differentially expressed across at least 2 strains. Subsequently, individual t-tests were used to compare expression levels in WT, OC, and PX in all 3 pairwise comparisons for each of these 188 genes. A significance cutoff value of P < 0.10 was used to approximately correspond to the 10% SAM false-positive rate. Fourteen of the genes selected by SAM were discarded after t-test analysis due to the absence of significant differences with P < 0.10. The remaining 174 genes were placed in 12 distinct expression profile categories. Because our analysis focuses on evolutionary changes in overall patterns of gene expression (rather than the magnitude or significance of changes in particular genes), individual P values for microarray data comparisons are not reported but are available upon request.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
The qRT-PCR Analysis
First, we used qRT-PCR to monitor development-specific gene expression in WT, OC, and PX of 5 genes that were previously shown to display major transcriptional increases during development in standard laboratory strains and that are essential for development in WT (sdeK, csgA, fruA, devR, and MXAN3227; listed in temporal order of induction, fig. 1). Each of these genes is turned on at a particular stage of development (ranging from 0 to 18 h after the onset of starvation) and their induction depends in different combinations on the intracellular stringent response signal (p)ppGpp and the intercellular A- and C-signals (fig. 1).
The sdeK gene encodes a histidine protein kinase necessary for initiation of cell aggregation (Garza et al. 1998; Pollack and Singer 2001). The expression of sdeK is induced by the stringent response immediately after cells are exposed to starvation (Kroos et al. 1986). The csgA gene encodes the precursor protein of the 17 kDa intercellular C-signal protein (Kim and Kaiser 1990; Lobedanz and Søgaard-Andersen 2003) that induces and coordinates aggregation and sporulation (Kim and Kaiser 1991; Li et al. 1992; Kruse et al. 2001). Expression of the csgA gene is induced after 3–6 h of starvation by the stringent response (Crawford and Shimkets 2000) and by C-signaling (Kim and Kaiser 1991). The fruA gene encodes a response regulator of 2-component signal transduction systems and is required for aggregation and sporulation (Ogawa et al. 1996; Ellehauge et al. 1998). The expression of fruA is induced after 4 h in response to the intercellular A-signal that communicates information about the density of starving cells, which must be sufficiently high for development to proceed (Kaplan and Plamann 1996). The devR gene encodes a protein required for sporulation and is induced after 6 h in a C-signal and FruA dependent manner (Kroos and Kaiser 1987; Viswanathan et al. 2007). Finally, the gene MXAN3227, which was originally defined by the Tn5lac 7536 insertion, is induced after 18 h in a devR dependent manner and is required for spore formation and spore coat formation (Licking et al. 2000).
Developmental expression of all 5 genes was found to be significantly higher in PX than in OC (fig. 3; supplementary table S1, Supplementary Material online). Four of the genes (sdeK, csgA, fruA, and MXAN3227) showed significantly elevated expression in PX relative to WT based on Tukey honestly significantly different (HSD) post hoc strain comparisons (Tukey 1949) after 2-way analysis of variance (ANOVA) was performed (supplementary table S1, Supplementary Material online). Although pairwise Tukey strain comparisons of devR expression at individual time points were not performed due to lack of an interaction between strain and time in the ANOVA (supplementary table S1, Supplementary Material online), devR nonetheless appears to show transiently elevated expression in PX relative to WT early in development (fig. 3). Developmental expression of sdeK and csgA did not differ between WT and OC but OC showed significantly reduced expression of fruA, devR, and MXAN3227 relative to WT. These experiments show that OC tends to show reduced or unchanged expression of several important developmental genes, whereas expression of these genes in PX is clearly elevated in all but one case.
|
Microarray Analysis
The qRT-PCR experiments suggested that the restored ability of PX to undergo development involves a novel transcription pattern compared with WT. To test this hypothesis more comprehensively, we carried out DNA microarray experiments using an array representing approximately 88% of M. xanthus genes (Jakobsen et al. 2004
; Overgaard et al. 2006). To analyze developmental gene expression in WT, OC, and PX, RNA was isolated from 5 time points throughout development (0, 6, 12, 18, and 24 h). In order to focus on major differences in gene expression levels between the different strains during development, we mixed RNA from the 6-, 12-, 18-, and 24-h time points in equal proportions to create pooled samples for each of the 3 strains. Thus, in these pooled samples, the amount of a specific mRNA species essentially represents an average of this mRNA throughout the first 24 h of development in that particular strain. As a common reference in all the microarray experiments, we used a sample consisting of equal amounts of RNA from WT from all 5 time points. This experimental design of comparing pooled samples from different time points is conservative with respect to detecting expression differences between strains because it preferentially reveals differences that are sustained over extended periods of development. This design also tends to mask small, transient differences and thus is expected to underestimate the total number of real expression differences among strains. In total, 174 genes were identified as being expressed differentially in either OC or PX (or both) relative to WT (supplementary table S2, Supplementary Material online) and were grouped into 12 distinct expression pattern categories (table 1). In total, 118 genes were found to be expressed differently in OC compared with WT (54 higher and 64 lower) (table 1, fig. 4). PX showed a greater number of differences from WT (144 total, 91 higher and 53 lower) than did OC. Finally, 162 of these genes (all categories except 5 and 9, table 1) are differentially expressed between OC and PX.
|
|
Among the 144 genes that are differentially expressed in PX compared with WT, 56 represent significant changes in PX from an unchanged level in OC (categories 1 and 2). The 31 genes in categories 8/9/12 and 5/6/7 were expressed at lower and higher levels, respectively, in both OC and PX compared with WT. Far more genes were upregulated in PX relative to WT than were downregulated (91 and 53, respectively), suggesting that expression changes causally responsible for the PX phenotype may be predominantly positive. Finally, the 57 genes in categories 3 and 11 have directionally opposite expression states in OC and PX (relative to WT). In contrast to the 144 genes just described, only 30 of the 118 genes with altered transcription in OC reverted to their ancestral transcriptional state during the evolution of PX (categories 4 and 10).
To compare the expression data obtained from the DNA microarray analysis with that from the qRT-PCR analysis, we analyzed the expression of 2 genes (sdeK and MXAN3277) included in the qRT-PCR analysis. The remaining 3 genes examined by qRT-PCR (csgA, fruA, and devR) either were not spotted on the microarrays or were represented by spots that were excluded from analysis due to poor quality. Consistent with qRT-PCR data (fig. 3A), the microarray data suggested greater expression of sdeK in PX than WT and OC (which had very similar expression level estimates in both the qRT-PCR and microarray experiments), but these differences were not significant (data not shown). The large increase in developmental expression of MXAN3227 in PX relative to WT and OC revealed by qRT-PCR analysis (fig. 3E) was also observed in the microarray experiment (P < 0.001 for both comparisons, data not shown). However, the decreased expression of MXAN3227 in OC late in development relative to WT (fig. 3E) was not detected in the microarray experiment, presumably due to the pooling of developmental time point samples. These results support the expectation that the design of the microarray experiment should result in an underestimation of the total number of significant expression differences among these strains.
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
Evolutionary restoration of major phenotypes might be caused at the sequence level by either direct mutational reversion to an ancestral state or novel compensatory mutations. Analogously, the transcriptome that underlies a reevolved phenotype might be largely similar to an ancestral state or may exhibit novel features. Here we have shown that an evolutionary restoration of developmental proficiency in M. xanthus was accompanied by a shift to a novel transcriptomic profile rather than by reversion back to the ancestral WT profile.
In particular, qRT-PCR experiments showed that 4 (sdeK, csgA, fruA, and MXAN3227) out of 5 examined genes essential for multicellular development in WT were persistently expressed at higher levels in the evolved strain with restored development (PX) than in the ancestral proficient strain (fig. 3). In contrast, developmental expression was lower in the intermediate, defective strain (OC) at 3 genes (fruA, devR, and MXAN3227) relative to WT and was lower in OC than in PX at all 5 genes. The observation of enhanced PX expression of genes known to be essential for development suggested that developmental expression of numerous other genes would also differ between these 2 strains. This hypothesis was confirmed by genome-wide transcriptional profiling using DNA microarrays. These analyses revealed a large (and likely underestimated) number of genes that are expressed differently in PX compared with WT during development and indicate that restored development in PX is based on a novel, rather than merely reverted, transcriptional program.
Our microarray data do not reveal which of the expression changes resulting from the evolution of OC into PX (162 detected here) play functional roles in the restoration of fruiting body formation and sporulation in PX. The genes with altered expression in PX perform a wide variety of functions (supplementary table S1, Supplementary Material online) and extensive genetic studies would be required to determine which differences between PX and WT are functionally significant and to characterize their corresponding mechanistic functions and regulatory relationships. However, fully 76% (132 genes) of detected changes from OC to PX represent evolutionarily novel expression states in PX that are significantly different not only from OC but also from WT. This result strongly suggests that a majority of those differences between OC and PX that are in fact functional with respect to restored development involve novel, rather than reverted, expression states in PX relative to WT.
Interestingly, the overall developmental gene expression profile of OC is more similar to that of WT (118 differences from WT) than is that of PX (144 differences from WT, fig. 4). PX and OC differ in the expression of 162 genes. Sequencing of the PX genome and subsequent analyses revealed that OC harbors 14 mutations relative to WT, whereas the PX phenotype was caused by only a single mutation. Our data indicate that the 14 mutations in OC combined cause fewer changes in developmental expression (118) than does the single PX mutation (162).
It is interesting to note that OC displays WT expression patterns of early developmental markers (sdeK and csgA), suggesting that the stringent response is unaffected in OC. The decreased expression of fruA in OC may reflect decreased production of A-signal in OC. This interpretation would be consistent with the observation that OC development is rescued by codevelopment with WT. The decreased expression of fruA would contribute to the reduced expression of devR and MXAN3227. The transformation of OC into PX was previously found to be caused by single mutation that altered the central position of a 7-base cytosine homopolymer located 128 bases upstream from the start codon of a predicted GNAT family acetyltransferase of unknown function (Fiegna et al. 2006; Velicer et al. 2006). This PX mutation may affect the expression of a regulatory gene or the function of a small regulatory RNA that acts early as well as late in development to cause the increased expression of sdeK, csgA, fruA, and MXAN3227. In all 4 cases, these increases led to expression levels in PX that were higher than WT, including 2 cases (sdeK and csgA) in which expression in OC is indistinguishable from that in WT (supplementary table S1, Supplementary Material online). This effect may be mediated by regulating the stringent response or a broad-spectrum transcriptional inducer that acts downstream from the stringent response but upstream from sdeK, csgA, fruA, and MXAN3227 in the M. xanthus developmental program. Alternatively, the PX mutation might modify the transcription of multiple developmentally regulated genes independently of one another. Sequence analysis suggests that the region in which the PX mutation lies may contain a small regulatory RNA gene (Yu Y-TN, unpublished data). Identification of genes regulated by such a putative regulatory RNA may shed light on the mechanistic significance of gene expression states unique to PX.
The compensatory mutation that restores development in PX represents a single adaptive event causing a large jump in biological complexity (Orr 2005) that is both social and developmental in nature. This mutation does not revert developmental transcription patterns back to their ancestral state. Rather, it pleiotropically enhances the transcription of several major developmental genes (relative to ancestral levels) that act at multiple different entry points and temporal stages of the genetic pathway underlying multicellular development in M. xanthus. Moreover, this mutation generated a novel overall transcriptional profile during development that is expected to include a large proportion of the gene expression changes that functionally contribute to the PX phenotype. The gene expression results presented here, combined with the previous observation that M. xanthus can evolve novel social swarming functions when the ancestral mechanism is disrupted (Velicer and Yu 2003), suggest that the reevolution of complex multicellular functions may commonly be accomplished by the generation of novel gene regulation states rather than reversion to ancestral states.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
Supplementary tables S1 and S2 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Footnotes |
|---|
1 Present address: Department of Ecology and Evolution, University of Chicago, Chicago IL 60637.
Jennifer Wernegreen, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionSupplementary MaterialReferences |
|---|
Burch CL, Chao L. Evolution by small steps and rugged landscapes in the RNA virus phi6. Genetics (1999) 151:921–927.
Cooper VS, Lenski RE. The population genetics of ecological specialization in evolving Escherichia coli populations. Nature (2000) 407:736–739.[CrossRef][Medline]
Crawford EW, Shimkets LJ. The Myxococcus xanthus socE and csgA genes are regulated by the stringent response. Mol Microbiol (2000) 37:788–799.[CrossRef][Web of Science][Medline]
Crill WD, Wichman HA, Bull JJ. Evolutionary reversals during viral adaptation to alternating hosts. Genetics (2000) 154:27–37.
Domingo E, Holland JJ. RNA virus mutations and fitness for survival. Annu Rev Microbiol (1997) 51:151–178.[CrossRef][Web of Science][Medline]
Ellehauge E, Nørregaard-Madsen M, Søgaard-Andersen L. The FruA signal transduction protein provides a checkpoint for the temporal co-ordination of intercellular signals in Myxococcus xanthus development. Mol Microbiol (1998) 30:807–817.[CrossRef][Web of Science][Medline]
Fiegna F, Velicer GJ. Competitive fates of bacterial social parasites: persistence and self-induced extinction of Myxococcus xanthus cheaters. Proc Roy Soc Biol Sci. (2003) 270:1527–1534.
Fiegna F, Yu Y-TN, Kadam SV, Velicer GJ. Evolution of an obligate social cheater to a superior cooperator. Nature (2006) 441:310–314.[CrossRef][Medline]
Garza AG, Pollack JS, Harris BZ, Lee A, Keseler IM, Licking EF, Singer M. SdeK is required for early fruiting body development in Myxococcus xanthus. J Bacteriol (1998) 180:4628–4637.
Harris BZ, Kaiser D, Singer M. The guanosine nucleotide (p)ppGpp initiates development and A-factor production in Myxococcus xanthus. Genes Dev (1998) 12:1022–1035.
Heineman RH, Molineux IJ, Bull JJ. Evolutionary robustness of an optimal phenotype: re-evolution of lysis in a bacteriophage deleted for its lysin gene. J Mol Evol (2005) 61:181–191.[CrossRef][Web of Science][Medline]
Hodgkin J, Kaiser D. Cell-to-cell stimulation of movement in nonmotile mutants of Myxococcus. Proc Natl Acad Sci USA (1977) 74:2938–2942.
Jakobsen JS, Jelsbak L, Jelsbak L, Welch RD, Cummings C, Goldman B, Stark E, Slater S, Kaiser D. 
54 enhancer binding proteins and Myxococcus xanthus fruiting body development. J Bacteriol (2004) 186:4361–4368.
Kaiser D. Signaling in myxobacteria. Annu Rev Microbiol (2004) 58:75–98.[CrossRef][Web of Science][Medline]
Kaplan HB, Plamann L. A Myxococcus xanthus cell density-sensing system required for multicellular development. FEMS Microbiol Lett (1996) 139:89–95.[Web of Science][Medline]
Kim SK, Kaiser D. C-factor: a cell-cell signaling protein required for fruiting body morphogenesis of M. xanthus. Cell (1990) 61:19–26.[CrossRef][Web of Science][Medline]
Kim SK, Kaiser D. C-factor has distinct aggregation and sporulation thresholds during Myxococcus development. J Bacteriol (1991) 173:1722–1728.
Kroos L, Kaiser D. Expression of many developmentally regulated genes in Myxococcus depends on a sequence of cell interactions. Genes Dev (1987) 1:840–854.
Kroos L, Kuspa A, Kaiser D. A global analysis of developmentally regulated genes in Myxococcus xanthus. Dev Biol (1986) 117:252–266.[CrossRef][Web of Science][Medline]
Kruse T, Lobedanz S, Berthelsen NMS, Søgaard-Andersen L. C-signal: a cell surface-associated morphogen that induces and coordinates multicellular fruiting body morphogenesis and sporulation in M. xanthus. Mol Microbiol (2001) 40:156–168.[CrossRef][Web of Science][Medline]
Kuner JM, Kaiser D. Fruiting body morphogenesis in submerged cultures of Myxococcus xanthus. J Bacteriol (1982) 151:458–461.
Levin BR, Perrot V, Walker N. Compensatory mutations, antibiotic resistance and the population genetics of adaptive evolution in bacteria. Genetics (2000) 154:985–997.
Li S, Lee BU, Shimkets LJ. csgA expression entrains Myxococcus xanthus development. Genes Dev (1992) 6:401–410.
Licking E, Gorski L, Kaiser D. A common step for changing cell shape in fruiting body and starvation-independent sporulation in Myxococcus xanthus. J Bacteriol (2000) 182:3553–3558.
Lobedanz S, Søgaard-Andersen L. Identification of the C-signal, a contact-dependent morphogen coordinating multiple developmental responses in Myxococcus xanthus. Genes Dev (2003) 17:2151–2161.
MacLean RC, Bell G, Rainey PB. The evolution of a pleiotropic fitness tradeoff in Pseudomonas fluorescens. Proc Natl Acad Sci USA (2004) 101:8072–8077.
Ogawa M, Fujitani S, Mao X, Inouye S, Komano T. FruA, a putative transcription factor essential for the development of Myxococus xanthus. Mol Microbiol (1996) 22:757–767.[CrossRef][Web of Science][Medline]
Orr HA. The genetic theory of adaptation: a brief history. Nat Rev Genet (2005) 6:119–127.[CrossRef][Web of Science][Medline]
Overgaard M, Wegener-Feldbrügge S, Søgaard-Andersen L. The orphan response regulator DigR is required for synthesis of extracellular matrix fibrils in Myxococcus xanthus. J Bacteriol (2006) 188:4384–4394.
Pollack JS, Singer M. SdeK, a histidine kinase required for Myxococcus xanthus development. J Bacteriol (2001) 183:3589–3596.
Shimkets LJ. Intercellular signaling during fruiting-body development of Myxococcus xanthus. Annu Rev Microbiol (1999) 53:525–549.[CrossRef][Web of Science][Medline]
Singer M, Kaiser D. Ectopic production of guanosine penta- and tetraphosphate can initiate early developmental gene expression in Myxococcus xanthus. Genes Dev (1995) 9:1633–1644.
Spiers AJ, Kahn SG, Bohannon J, Travisano M, Rainey PB. Adaptive divergence in experimental populations of Pseudomonas fluorescens. I. Genetic and phenotypic bases of wrinkly spreader fitness. Genetics (2002) 161:33–46.
Teotonio H, Rose MR. Perspective: reverse evolution. Evolution (2001) 55:653–660.[CrossRef][Web of Science][Medline]
Tukey JW. Comparing individual means in the analysis of variance. Biometrics (1949) 5:99–114.[Medline]
Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA (2001) 98:5116–5121.
Velicer GJ, Kroos L, Lenski RE. Loss of social behaviors by Myxococcus xanthus during evolution in an unstructured habitat. Proc Natl Acad Sci USA (1998) 95:12376–12380.
Velicer GJ, Kroos L, Lenski RE. Developmental cheating in the social bacterium Myxococcus xanthus. Nature (2000) 404:598–601.[CrossRef][Medline]
Velicer GJ, Raddatz G, Keller H, Deiss S, Lanz C, Dinkelacker I, Schuster SC. Comprehensive mutation identification in an evolved bacterial cooperator and its cheating ancestor. Proc Natl Acad Sci USA (2006) 103:8107–8112.
Velicer GJ, Yu YT. Evolution of novel cooperative swarming in the bacterium Myxococcus xanthus. Nature (2003) 425:75–78.[CrossRef][Medline]
Viswanathan P, Ueki T, Inouye S, Kroos L. Combinatorial regulation of genes essential for Myxococcus xanthus development involves a response regulator and a LysR-type regulator. Proc Natl Acad Sci USA (2007) 104:7969–7974.
Wall D, Kolenbrander PE, Kaiser D. The Myxococcus xanthus pilQ (sglA) gene encodes a secretin homolog required for type IV pilus biogenesis, social motility, and development. J Bacteriol (1999) 181:24–33.
Whiting MF, Bradler S, Maxwell T. Loss and recovery of wings in stick insects. Nature (2003) 421:264–267.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||