planctomycetes comparative genomics - american society for

1

Planctomycetes Comparative Genomics: Identification 1 of Proteins Likely Involved in Morphogenesis, Cell 2

Division and Signal Transduction 3 4 Christian Jogler 1,2,*), Jost Waldmann 3), Xiaoluo Huang 4), Mareike Jogler 2), Frank 5 Oliver Glöckner3), Thorsten Mascher 4), Roberto Kolter 1) 6 1) Harvard Medical School, Boston 7 2) DSMZ, Brunswick 8 3) MPI for Marine Microbiology Bremen, & Jacobs University Bremen 9 4) LMU, Munich 10 *) corresponding author 11

Copyright © 2012, American Society for Microbiology. All Rights Reserved.J. Bacteriol. doi:10.1128/JB.01325-12 JB Accepts, published online ahead of print on 21 September 2012

on April 4, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

2

Abstract 12 Members of the Planctomycetes clade share many unusual features for bacteria. Their 13 cytoplasm contains membrane-bound compartments, they lack peptidoglycan and FtsZ, 14 divide by polar budding, and they are capable of endocytosis. Planctomycete genomes 15 have remained enigmatic, generally being quite large (up to 9 megabases) and, on 16 average, 55% of their predicted proteins are of unknown function. Importantly, proteins 17 related to the Planctomycetes' unusual traits remain largely unknown. Thus, we embarked 18 on bioinformatic analyses of these genomes in an effort to predict proteins that are likely 19 to be involved in compartmentalization, cell division and signal transduction. We used 20 three complementary strategies. First, we defined the Planctomycetes core-genome and 21 subtracted genes of well-studied model organisms. Second, we analyzed gene content and 22 synteny of morphogenesis and cell division genes and combined both methods using a 23 “guilt-by-association” approach. Third, we identified signal transduction systems as well 24 as sigma factors. These analyses provide a manageable list of candidate genes for future 25 genetic studies and provide evidence for complex signaling in the Planctomycetes akin to 26 that observed in bacteria with complex life styles such as Myxococcus xanthus. 27


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

3

Introduction 28 In many respects the Planctomycetes are extremely unusual bacteria (12). There is still 29 controversy among evolutionary biologists regarding Planctomycetes' phylogeny. They 30 have been proposed to be the deepest branching bacterial phylum (4, 23) or the most 31 rapidly evolving bacteria (62). Their remote relatedness to Chlamydia (55, 61) has led to 32 a broadly accepted hypothesis that there is a superphylum consisting of the 33 Planctomycetes, Verrucomicrobia, Chlamydia (PVC) (46, 59). Regardless of their exact 34 phylogeny, all Planctomycetes share a very unusual cell plan that differs dramatically 35 from the Gram-positive and Gram-negative bacteria (13, 28). They lack peptidoglycan; 36 instead they have a proteinaceous cell wall, a fact that renders them resistant against β-37 Lactam antibiotics (6, 26, 27). Underneath the proteinaceous cell wall lays the 38 cytoplasmic membrane (CM). Strikingly, the cytoplasm of planctomycetes is divided by 39 an intracytoplasmic membrane (ICM) into the paryphoplasm and the pirellulosome or 40 riboplasm (28). The DNA is highly condensed and forms a nucleoid that occupies only a 41 fraction of the pirellulosome (12). Interestingly, in Gemmata obscuriglobus the nucleoid 42 is surrounded by two closely opposed double membranes thus resembling the enveloped 43 nucleus of a eukaryotic cell (12, 13). But in contrast to eukaryotes, this nucleus-like 44 structure contains ribosomes. However, since most ribosomes are situated within the 45 pirellulosome, many mRNAs generated in the nucleus-like compartment of G. 46 obscuriglobus must somehow be transported through both double membranes to be 47 translated. Yet, nothing is known about such transfer mechanisms and whether they are 48 related to eukaryotic mRNA transport (12). Planctomycetes are also unusual in that their 49 membranes contain sterols (42). Further, Planctomycetes and some other members of the 50 PVC-superphylum are the only organisms among bacteria and archaea that encode 51 proteins with high structural similarity to eukaryotic membrane coat proteins (MC), that 52 play a major role in the formation of coated vesicles. The planctomycetal MC-like 53 proteins have β-propeller domains followed by stacked pairs of α–helices (SPAH) as do 54 their eukaryotic counterparts. In addition, a significant amount of the MC-like proteins 55 are localized in close proximity to the ICM or to vesicles within the paryphoplasm (48). 56 Such vesicles have been shown to allow endocytosis-like uptake of proteins into the 57


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

4

paryphoplasm of G. obscuriglobus cells (29), a trait long believed to be a hallmark of 58 eukaryotic cells (21). In contrast to all other bacteria, the polar planctomycetes divide via 59 budding, in an FtsZ-independent manner. As a consequence their cell division resembles 60 more that of budding yeast than the division of the typical bacterium that relies on FtsZ. 61 Thus, planctomycetes challenge the prokaryote/eukaryote dichotomy in several ways (10). 62 In the face of these unusual features, one might expect that planctomycetal genomes 63 might contain many Archaea-like or Eukarya-like genes. However, this seems not to be 64 the case (1, 11, 55). Thus, we embarked on planctomycetes comparative genomics with 65 the aim of identifying proteins likely related to their unusual biology. The results of these 66 analyses now provide candidate proteins for performing subsequent genetic analyses. 67 Given that genetic tools are only available for P. limnophilus, we mainly focused on this 68 model organism. 69


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

5

Material and Methods 70 Phylogenetic analyses of Planctomycetes 71 Aligned 16S rRNA gene sequences were derived from the SILVA database 72 (http://www.arb-silva.de/download/arb-files, release 108 (47)) and the alignment was 73 corrected manually. Phylogenetic trees were calculated employing the ARB software 74 package (31). The RAxML module (rate distribution model GTRGAMMA; rapid 75 bootstrap analysis algorithm) performed the Maximum Likelihood analysis while 76 Neighbor Joining trees were calculated with the ARB Neighbor Joining tool using the 77 Felsenstein correction. Bootstrap values were determined using 1000-fold resampling 78 (31). Accession numbers of sequences used are listed in Tab. S1. 79 80 Distilling the planctomycetal core genome 81 First, an NCBI BLAST database (‘core-contributors’) was generated containing all 82 protein sequences of all available Planctomycetes (Tab. 1), preserving their taxonomic 83 information. Second, an all-against-all BLASTP search was used to extract all homologs 84 within the dataset. To exclude matches to the originating genome, the mask option (-85 negative_gilist) of BLAST was used. If a protein matched the same database entry 86 multiple times only the best BLAST-match was considered. Hits with query coverage 87 higher than 60% and below the E-value cutoff 1e-5 were taken into account. Additionally, 88 the number of species in the associated list of BLAST matches was counted for each 89 query protein. Queries not matching all taxons in the database, except the originating one, 90 were discarded. Subsequently, sequences of queries fulfilling selection criteria were 91 compiled to a list and all genes among the list linked by reciprocal BLAST-hits were 92 clustered. If a BLAST match did not link back to the query the corresponding sequence 93 was rejected. Consequently each cluster represents a group of homologous genes, with a 94 high probability for sharing the same function. 95 Trimming the planctomycetal core genome using in silico subtraction 96 We compiled a set of genes from the well-studied model organisms (E. coli str. K-12 97 substr. MG1655, accession: U00096 and B. subtilis subsp. Subtilis str. 168, accession: 98 AL009126) and a second BLAST database ('blacklist') was generated out of those. All 99


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

6

genes of the planctomycetes core-genome were compared against this 'blacklist' database 100 using BLAST. Genes reaching the E-value cutoff 1e-5 and a minimum coverage of 65% 101 led to the deletion of the entire cluster from the planctomycetal core genome set. 102 Corresponding Perl scripts are available upon request. A list of all protein sequences and 103 the resulting clusters is available in the supplemental section (Tab. S2). 104 Analysis of the Planctomyces limnophilus core genome 105 Since P. limnophilus is the only planctomycete that can be genetically manipulated, we 106 focused on its core genome in order to identify candidate proteins for subsequent genetic 107 experiments. Consequently all sequences except those of P. limnophilus origin were 108 deleted from the 114 core-genome clusters obtained after in silico subtraction of E. coli 109 and B. subtilis. The resulting clusters were manually inspected and proteins likely 110 involved in cell division, shape determination or compartmentalization were selected as 111 starting points for a “guilt-by-association approach” (Tab. S3). The genomic context of 112 each selected gene was manually inspected to identify putative operons harboring at least 113 two members of the core genome. Five putative operons fulfilled these criteria and were 114 manually analyzed using BLAST, the NCBI database and the Conserved Domain 115 Database (CDD) for the functional annotation (36). 116 Cell shape and cell division proteins 117 Screening the literature, we composed a list of 64 proteins involved in cell division, 118 filament formation and cell shape determination from various bacteria, archaea and yeast 119 (Tab. S4). The query protein was compared against the NCBI database using BLAST 120 with default settings and “Planctomycetes” as taxa filter. Putative homologues were 121 reverse analyzed against the database. If the reverse BLAST experiment revealed proteins 122 from the same family of the original query, they are highlighted in green in Tab. S4. If 123 not, they are highlighted in yellow and referred to as “-like” proteins. Protein IDs, E-124 values and % identity are given in Tab. S4. 125 Gene synteny and content of dcw operon genes among Planctomycetes 126 Based on previous studies, we used planctomycetes' proteins homologous to those of E. 127 coli encoded within the dcw operon as a starting point (Tab. S4) (39, 46). Subsequently, 128


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

7

we manually determined the synteny of corresponding genes and visualized the gene-129 order (Fig. 4), while domains were determined as described above. 130 Analysis of planctomycetal signal transduction 131 The Microbial Signal Transduction database MiST2 (http://mistdb.com) was employed to 132 gain an overview of regulatory proteins encoded in the eight planctomycetal genomes 133 (57). ECF sequences were retrieved from the MiST database, cross-checked by an ECF-134 finder analysis (http://ecf.g2l.bio.uni-goettingen.de:8080/ECFfinder/), and all positive 135 sequences were added to a comprehensive ECF-dataset that contains a total of 362 136 sequences (Tab. S5). Subsequently, the protein sequences were aligned using the 137 ClustalW algorithm at EBI (http://www.ebi.ac.uk/). After manually removing gaps and 138 N- or C-terminal extensions, phylogenetic distances were calculated for the resulting 139 multiple sequence alignment files of the ECF core proteins employing the Fitch 140 Margoliash and least squares distances algorithms (9) as implemented in the BioEdit 141 sequence alignment editor (18). Individual branches of the tree were subjected to detailed 142 follow-up analyses as described (53) to identify potential anti-σ factors and/or additional 143 associated proteins. These analyses are based on genomic context conservation as a 144 measure for a potential functional link between their respective gene products. 145 on A

pril 4, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

http://jb.asm.org/

8

Results and Discussion 146 The planctomycetal core genome 147 The Planctomycetes clade displays enormous phylogenetic depth and its phylogeny is 148 still controversial (23, 46, 59). Thus, prior to determining the planctomycete core-genome, 149 we used 16S rRNA gene analysis to decide which species to include in our study for the 150 practical purpose of determining a group of organisms sufficiently related to each other to 151 produce a meaningful core genome. The maximum likelihood 16S rRNA gene tree shows, 152 with excellent bootstrap support (100), two distinct branches. One branch comprised 153 Annamox-planctomycetes while the other branch contained all other planctomycetal 154 species cultivated thus far (Fig. 1). This result provided us with a phylogenetic criterion 155 to leave Anammox-bacteria out of our analysis. Within this manuscript we thus 156 differentiate between “Planctomycetes” and “Anammox-bacteria”. However, whether or 157 not this has taxonomic implications is beyond the scope of this study. 158 Despite phylogenetic controversies, Planctomycetes are members of the domain Bacteria 159 and as such would be expected to divide using FtsZ. However FtsZ is absent in 160 planctomycetal genomes (2). In order to identify proteins that might be involved in cell 161 division, shape determination and compartmentalization we performed two independent 162 lines of analysis. Our planctomycetal core genome approach led to 114 predicted protein 163 clusters, containing 2908 proteins from all eight analyzed planctomycetes after in silico 164 subtraction of E. coli and B. subtilis genomes (Fig. 2). Each cluster contained between 8 165 and 764 proteins with an average of 25. This subtraction yielded important insights since 166 planctomycetes are difficult to cultivate and genetic tools are only available for P. 167 limnophilus. Since P. limnophilus is the only member of this phylum that has been 168 genetically modified, we focused on its core genome for subsequent analysis. P. 169 limnophilus contributed 274 proteins to the 114 core-clusters resulting in an average of 170 about 2 proteins per cluster. However, even in this model planctomycete mutant 171 construction is rather slow. Thus, we wanted to further reduce the number of candidate 172 genes for subsequent mutational studies. Consequently all P. limnophilus core proteins 173 were manually evaluated and six clusters were selected for in-depth analyses based on 174 their putative cell biological relevance (Tab. S3). One of these clusters, no. 23, contained 175


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

9

39 MC-like proteins encoded by all eight planctomycetal genomes under study and five 176 proteins belong to P. limnophilus. Despite MC-like proteins, no “smoking gun” was 177 found and given that most proteins are of unknown function we followed a “guilt-by-178 association" approach to identify putative targets. First, we manually inspected the 179 neighborhood within the six selected P. limnophilus core genome clusters. Second, we 180 focused on genomic regions in which at least two core genes were situated within a 181 putative operon. Five such putative operons (PO) fulfilled these criteria (Fig. 3; PO1-5), 182 while MC-like proteins were not among them. However, the five putative operons encode 183 proteins containing putative cell division related domains (e.g. von Willebrand factor, 184 PDZ and ATPase domains). Such domains have been found to be associated with 185 putative novel cell division mechanisms in Archaea (34, 35). Individually, each PO has 186 features that render it attractive for further study. 187 In PO1 we identified FHA, SpoIIE and LpxE domains. Such domains might be involved 188 in cell division and membrane organization, as FHA domains play a critical role in 189 hyphal fusion of Neurospora crassa while SpoIIE is required for asymmetric cell division 190 during sporulation in Bacillus (37). LpxE domains are known to selectively 191 dephosphorylate lipid A molecules which are present in the outer membrane of Gram-192 negative bacteria (20). This finding is unexpected, since planctomycetes lack such an 193 outer membrane. However, further genes required for the biosynthesis of lipid A were 194 found in R. baltica indicating that planctomycetes might produce lipid A (17, 24). 195 PO2 encodes a membrane-bound dehydrogenase and heme-binding domains. This 196 configuration is frequently found in Planctomycetes and Verrucomicrobia but is not 197 found in any other bacteria sequenced thus far. Furthermore, dihydrolipoamide 198 dehydrogenase and glycine cleavage H-protein domains were identified. Most strikingly, 199 some similarity to the MinD superfamily "P-loop ATPase ferredoxin" domains could be 200 detected. MinD participates in the spatial regulation of the formation of the cytokinetic 201 FtsZ ring and thus has a major function during cell division (41). 202 The third identified operon, PO3 encodes proteins which appear related to signal 203 transduction. Two of them contain von Willebrand factor type A domains. Thus, they 204 might be involved in processes such as membrane modeling, cellular differentiation and 205


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

10

adhesion. Genes encoding proteins of unknown function, such as Plim_0822 might be 206 responsible for planctomycetes-specific traits and their being near the "usual suspects" 207 such as Willebrand factor domains makes them interesting targets for subsequent genetic 208 experiments. 209 PO4 encodes 12 proteins. While two might be related to sugar metabolism, four are 210 hypothetical or contain domains of unknown function. Thus, mainly the vWF and 211 ATPase domains point towards the involvement of this putative operon in cell division or 212 membrane remodeling. Interestingly, Plim_1419 might resemble a gas vesicle-related 213 protein, GvpN. However, gvpN deletion mutants of Halobacterium salinarium revealed 214 that GvpN is not essential for gas vesicle assembly but may enhance their synthesis in 215 some way (40). 216 PO5 is the largest putative operon and its component proteins contain several ATPase 217 domains. In addition PDZ domains that may mediate protein-protein interactions are 218 found together with trypsin domains indicating putative proteases. Furthermore a 219 bacterial cell adhesion CARDB domain could be found. In addition, the fact that three 220 core genome members (Plim_1802-1804) are next to each other makes PO5 a promising 221 candidate for future investigation. For a more detailed description of the five putative 222 operons, see supporting text. 223 As expected, the comparative genomic approach revealed several target proteins from P. 224 limnophilus for future genetic analyses. Interestingly, while they are part of the core 225 genome, MC-like protein encoding genes show a patchy distribution in the P. limnophilus 226 genome and do not cluster with other core genome members. However, such proteins 227 were recently demonstrated to be involved in endocytosis of proteins in G. obscuriglobus 228 and are thus without doubt worth studying in detail (29, 48). 229 230 Cell division and cell shape determination 231 All eight planctomycete genomes used in this study were screened for genes encoding 232 one of 64 cell-morphogenesis or -division related proteins that were described in different 233 bacteria, archaea or yeast, with the results summarized in Table S4. Table 2 provides a 234


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

11

selection of proteins that were found in at least one planctomycete. As shape-determining 235 proteins we included representatives from all bacterial actin-like families, ParM, FtsA, 236 AlfA, Alp7 and MreB (49). The gene encoding the bacterial actin homolog MreB was 237 found in Planctomyces species and in B. marina. However, we failed to identify any 238 bacterial tubulin-like protein using BtubA and B as well as FtsZ as a query (32, 45). 239 Recently, a new family of FtsZ-like proteins (FtsZl-1) was described and B. marina, G. 240 obscuriglobus, P. staleyi and R. baltica were shown to encode such a protein (34). We 241 failed to detect homologues to such FtsZl-1 proteins in all other sequenced 242 planctomycetes using almost the same approach that led to the discovery of FtsZl-1 (34). 243 In addition, FtsZl-1 proteins lack the T6-H7 structure and thus might not be able to form 244 filaments at all (34). Thus, the FtsZl-1 might very well fulfill a function other than Z-ring 245 formation during cell division in planctomycetes. In addition, we screened for proteins 246 involved in archaeal cell division. Two FtsZ-independent division mechanisms are 247 known in the domain Archaea (35). One is homologous to the eukaryotic ERCRT-III 248 system and requires Snf and VPS4 proteins. While we could not detect the presence of 249 Snf encoding genes in planctomycetal genomes using BLAST (see Tab. S4), VPS4 250 revealed several putative homologs. However a critical diagnostic feature - the amino-251 terminal microtubule interaction and transport (MIT) domain - present in all archaeal 252 VPS4 proteins is missing in their putative planctomycetal counterparts. Thus, we 253 conclude that planctomycetes lack VPS4. The third archaeal division mechanism is still 254 puzzling but seems to involve actin-like proteins. As mentioned above, by our search 255 approach MreB homologues could only be found in four planctomycetes and can thus not 256 explain a universal mechanism of planctomycetal cell division. Finally, we did not detect 257 homologues of the protein putatively involved in Z-ring formation in Anammox-bacteria, 258 kustd_1438 (58). The absence of this protein among planctomycetes other than 259 Anammox-bacteria further supports our initial assumption to focus on one part of the 260 planctomycetal order as two independent cell division mechanisms might further support 261 the differences between both lineages. Since cytokinesis of planctomycetes parallels the 262 division of yeast, we screened for septin proteins known to be crucial for septum 263 formation in Saccharomyces (3, 43). However, we did not identify any homologs among 264 planctomycetes using the basic BLAST algorithm. 265


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

12

In contrast to putative Z-ring forming proteins, FtsW was found in Planctomyces species, 266 while genes for other cell division related proteins, such as FtsE+K, ParA, CpaE, EnvA 267 and MraW+Y could be detected in all planctomycetal genomes. However, MraZ was 268 only found in Planctomyces and Blastopirellula species. Besides cell division, 269 'peptidoglycan related' proteins such as MurC, D, E, F and G as well as Ddl and FtsI 270 showed a more complex distribution pattern. While all proteins are encoded in 271 Planctomyces species, all except MurE and Ddl are absent in all other planctomycetes. 272 MurE was found besides Planctomyces in B. marina, R. baltica and P. staleyi, while Ddl 273 was absent in B. marina (Tab. 2). 274 Beyond gene content we investigated the gene synteny of cell morphogenesis and 275 division genes. Here our starting point was the dcw operon structure of E. coli as a 276 comparison. The dcw operon is highly conserved among 'rod-shaped' bacteria and 277 “genomic channeling” has been proposed to force colocalization of genes involved in cell 278 division and peptidoglycan synthesis to optimize septum formation (39). While genomic 279 channeling nicely explained the persistence of the dcw operon over a broad taxonomic 280 range (39) (Fig. 4), this mechanism is supposed to be restricted to 'rod-shaped' cells and 281 transitions to other morphologies seem to involve genomic rearrangements and loss of 282 gene-order (54). This was supported by the first investigation of planctomycetal dcw-283 gene clusters (46). Despite the few planctomycetal genomes available at this time the 284 conclusion was drawn that the last planctomycetal ancestor (LPA) contained a complete 285 dcw operon and gene content and order was gradually lost (46). This hypothesis is 286 supported by our data as shown in Table 2 and Fig. 4: All three Planctomyces species 287 contain most dcw operon genes found in all planctomycetes. However, gene synteny is 288 less conserved among Planctomyces than gene content. While P. maris has four putative 289 operons and one single murD-like gene, the dcw operon members of E. coli are scattered 290 among eight different putative operons and one solitary murD in P. brasiliensis. In P. 291 limnophilus the situation is even more complex, as dcw genes are localized across nine 292 different putative operons. Among all other planctomycetes, P. staleyi and R. baltica 293 contain four dcw-operon genes while B. marina, G. obscuriglobus and I. pallida harbor 294 three, none of which colocalize. Thus, gene synteny is less conserved then gene content 295 in regard to planctomycetal dcw genes, while the gradual loss becomes more obvious 296


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

13

since our study used more than double the number of planctomycetal genomes than 297 previous analysis (46). Interestingly, gene content and synteny reflect the 16S rRNA gene 298 tree phylogeny (Fig. 1). However, the presence of peptidoglycan (PG) related genes is 299 difficult to explain as planctomycetes are known to lack a PG cell wall (12) and genes 300 without function tend to erode quickly. For PG-wall less Chlamydia, cell division has 301 been shown to be altered by beta-lactam antibiotics. Consequently speculations of an 302 involvement of PG in Z-ring replacement arose (2). However, consistent with earlier 303 studies we failed to detect any growth inhibition caused by beta-lactam antibiotics such 304 as ampicillin among planctomycetes thus far (6, 22, 26). 305 Taken together, our finding that several 'cell division' and 'peptidoglycan' associated 306 proteins known from bacteria are encoded by some, if not all Planctomycetes (Tab. 2) 307 might suggest that their last ancestor might have divided in an FtsZ-dependent manner 308 and was confined by a PG layer. 309 310 Putative cell-division related operons 311 Following a “guilt-by-association” approach, we screened putative planctomycetal dcw 312 gene containing operons for the presence of core-genome members. The rationale 313 driving this analysis is the assumption that cell-division-related dcw operon genes might 314 be found near genes involved in the FtsZ-independent division of planctomycetes. Genes 315 encoding FtsZ-substituting proteins most likely arose in the last planctomycetal ancestor 316 (LPA) and would consequently be absent from genomes of other bacteria such as E. coli 317 or B. subtilis. Such planctomycetal cell division genes are likely to be conserved among 318 all planctomycetes and to be part of their core genome. However, only two putative 319 operons fulfilled our screening criteria: One was found within the genome of P. 320 brasiliensis, spanning 25 kb and containing 14 genes, of which two are dcw operon 321 related (ftsI and mraW), while four are members of the planctomycetal core genome 322 (Plabr_0294+ 0295, Plabr_0300+ 0301) (Fig. 5). The second putative operon found in the 323 genome of I. pallida had a size of more than 38 kb and contained 26 genes. Beside of the 324 marY-like dcw gene, three planctomycetal core genome members, Isop_3026, 3036 and 325 3037, were identified (Fig. 5). However, since both organisms are not genetically 326


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

14

manipulable we performed a second screening, comparing the identified core genome 327 members putatively related to cell division from I. pallida and P. brasiliensis operons 328 against the genome of P. limnophilus (Plim). This revealed two and three putative Plim 329 operons, respectively (Fig. 5; for detailed description of individual genes and proteins see 330 supporting text). 331 Many proteins encoded by the genes of PO6-10 contained domains putatively related to 332 cell division such as the ATPase- or to chromosome segregation domains. In addition, 333 many genes contain domains of unknown function that are exclusively found in 334 Planctomycetes or Verrucomicrobia (Fig. 5). Most strikingly, three of four putative 335 operons identified following the “guilt-by-association” approach turned out to contain 336 two or three genes that are part of the planctomycetal core genome. In the putative 337 operons six and seven (PO6+7), two core genome members with identical domain 338 architecture colocalize (Plim_2793+2794 and Plim_0775+0776 respectively). 339 Interestingly the same domain architecture and localization pattern was observed in two 340 areas of a P. brasiliensis putative operon, where the first two genes, Plabr_0294+0295 341 colocalize with ftsI and the other two Plabr_0300+0301 with mraW. This finding makes 342 Plim putative operons six and seven interesting candidates for future genetic studies. 343 All ten identified putative operons (Fig. 3 and 5) provide some indications that their gene 344 products are involved in cell shape determination, membrane organization, chromosome 345 segregation or cell division. The planctomycetal proteins of unknown function, 346 containing known domains or not, might be the most interesting ones as new biological 347 functions might await elucidation. The rationale for our study was the in silico 348 identification of putative targets for subsequent genetic experiments and all putative 349 operons identified need future experimental verification to reveal their putative relevance 350 for planctomycetal key-traits. Wet lab experiments must now be carried out to determine 351 whether colocalized genes are co-transcribed or not. Experimentally verified operons 352 from the gene clusters identified in this study should then be subjected to subsequent 353 mutagenesis in P. limnophilus. Such experimental approaches should help to ultimately 354 identify the function of genes involved in planctomycete-specific traits. 355 356


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

15

Planctomycetal signal transduction and gene regulation 357 Complex life styles and cellular differentiation cycles necessitate the presence of equally 358 complex signaling cascades and regulatory networks to orchestrate and coordinate the 359 corresponding massive spatio-temporal modulations of gene expression patterns, as has 360 been observed in model organisms such as B. subtilis, C. crescentus or S. coelicolor (25, 361 30, 38). Understanding the regulatory and signaling potential embedded in genome 362 sequences of Planctomycetes species is therefore a prerequisite to unravel differentiation 363 processes in this phylum. We therefore decided to closely examine the repertoire of 364 signal transducing proteins of the planctomycetes species. 365 Based on mechanistic principles and phylogenetic relationship, three fundamental 366 mechanisms of bacterial signal transduction can be distinguished: one-component 367 systems (1CSs), two-component systems (2CSs) and extracytoplasmic function sigma 368 factors (ECFs) (52, 53, 56). Based on data from well-represented bacterial phyla such as 369 the Proteobacteria or the Firmicutes, an ‘average’ genome harbors about four times more 370 1CSs than 2CSs, and four times more of the latter than ECFs. But while the proportion of 371 1CSs and 2CSs is relatively constant for most phyla (56), the number of ECFs seems to 372 be much more variable amongst different species (52). With very few exceptions, the 373 absolute numbers of signaling proteins directly correlates with the size of the genome 374 (56). The distribution of planctomycetal signaling proteins supports both, the correlation 375 between genome size and number of signaling proteins as well as the ratio between 1CSs 376 and 2CSs (Table 3). In contrast, all planctomycetes are particularly rich in ECFs, which 377 equal or sometimes even outnumber the 2CSs, indicating an important role of these 378 proteins in the regulatory processes of this phylum. 379 One-component systems (1CSs) 380 1CSs are proteins that connect an input with an output on a single polypeptide chain. The 381 majority (about 85%) of 1CSs from most bacteria harbor a DNA-binding output domain, 382 as is the case for paradigms such as LacI or TetR. The output of the remaining 15% is 383 mostly realized using protein kinases/phosphatases or regulators involved in the making 384 and breaking of secondary messengers, such as adenylate/di-guanylate cyclases (56). In 385 planctomycetes, the percentage of DNA-binding 1CSs is much lower while 20-45% of all 386


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

16

1CSs are Ser/Thr protein kinases (STPKs) (Table 3). Such proteins, which are common 387 among eukaryotes, normally play a minor role in bacterial signal transduction and their 388 physiological role has only very recently been elucidated for few examples, most of 389 which are involved in differentiation or developmental processes (44). The reason for the 390 very different abundance of STPKs between prokaryotes and eukaryotes might reside in 391 the fact that in bacteria, with their chromosome located in the cytoplasm, most adaptation 392 and differentiation processes are directly regulated via differential transcription. For this, 393 bacteria predominantly use sensory histidine kinases (HKs) as part of 2CSs to activate 394 their cognate response regulators (RRs), which – once activated – mostly function as 395 DNA-binding proteins. In contrast, the eukaryotic chromosomes are spatially separated 396 from the cytoplasm in the nucleus. Hence, direct protein modifications (including 397 phosphorylations) play a much more important role in the cellular physiology and 398 differentiation. Thus, eukaryotes require larger numbers of protein kinases to modulate 399 the protein phosphorylation pattern in response to diverse stimuli. Considering that the 400 chromosome of e.g. Gemmata obscuriglobus is surrounded by two nuclear membranes, 401 the high abundance of STPKs might point towards a similar role of protein 402 phosphorylation in the differentiation processes of planctomycetes (12). However, this 403 does not imply any necessary evolutionary homology but merely a convergence of signal 404 transduction system characteristics due to analogies in cell organization. 405 Two-component systems (2CSs) 406 2CSs consist of a sensor protein, the histidine kinase (HK), which in the presence of a 407 suitable stimulus autophosphorylates. Subsequently a cognate partner protein, the 408 response regulator (RR), becomes active through a phosphoryl group transfer from the 409 HisKA domain of the HK to the receiver domain of the RR (15). This phosphorylation 410 then activates the output domain of the RR, which in most cases (about 85%) binds DNA 411 to orchestrate differential expression of its target genes (14, 56). Usually, the two partner 412 proteins are encoded by neighboring genes that are organized in an operon. In 413 planctomycetal genomes, the absolute number of 2CS proteins correlates well with 414 genomes sizes, and the ratio of 1CS to 2CS is comparable to other species (Table 3). 415 However, RRs outnumber HKs and only 50-70% of the RRs are DNA-binding proteins 416 (bacterial average 85%). Most planctomycetal RRs solely consist of a receiver domain 417


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

17

only and lack obvious output domains. Many planctomycetal 2CS sensor proteins 418 represent complex (hybrid) HKs, containing multiple putative input domains and often 419 more than one HisKA domain. Hybrid HK are part of complex phosphorelays, involving 420 the integration / modulation of the HK activity itself or the activity of downstream RRs 421 (7). The majority (about 70-80%) of HKs and RRs are encoded by orphan genes 422 unrelated to their respective partner, pointing towards a more complex mode of 2CS-423 dependent signal transduction in planctomycetes. Among such genes, one was found to 424 encode a novel type of RR, conserved in all eight planctomycetal genomes, but restricted 425 to this phylum. Its domain architecture consists of an N-terminal Ser/Thr/Tyr protein 426 kinase- (Pfam: Pkinase) and a C-terminal receiver (REC) domain (Fig. 6A), which is 427 normally found in the N-terminus. 428 The widely distributed domain architecture of RRs solely consisting of a receiver domain 429 might point towards a role of these proteins in shuttling phosphoryl-groups between 430 histidine kinases and histidine-containing phosphotransfer proteins (containing a 431 Pfam:HPt domain) within complex 2CS-based phosphorelays involved in integrating 432 different stimuli, as described the decision making process in B. subtilis sporulation or 433 the Rcs phosphorelay of Enterobacteria (8, 33, 60). This hypothesis seems attractive in 434 light of the higher abundance of RRs relative to HKs. A second explanation could be that 435 such ‘receiver-only’ RRs mediate their output through direct protein-protein interactions, 436 as most classically exemplified by the CheY-like RRs involved in the chemotaxis 437 systems discussed below. Taken together, only a small fraction of planctomycetal 2CSs 438 provide a simple one-to-one connection between a stimulus input and a cellular response 439 in the form of differential gene expression. Instead, the majority of 2CS proteins seem to 440 be involved in complex, highly interconnected multi-step phosphorelays required to 441 orchestrate some aspects of the complex life cycle described for this phylum. 442 Interestingly, the overall situation is very reminiscent of the Deltaproteobacterium 443 Myxococcus xanthus (50). However, the unusual domain configuration of some 444 planctomycetal RRs (Fig. 6A) remains enigmatic. Probably upon phosphorylation of the 445 receiver domain by a cognate HK, the N-terminal protein kinase domain is activated to 446 phosphorylate its target protein(s) as output of a 2CS-mediated signal transduction. While 447 both the physiological role and the signaling mechanism remains elusive, the 448


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

18

conservation of these novel RRs within all planctomycetal species and its absence outside 449 this phylum makes these proteins very interesting candidates for future functional studies. 450 Chemotaxis 451 While bacterial chemotaxis is derived from 2CSs, we treat such proteins as a separate 452 entity, because of the large number of unique domains and proteins dedicated to 453 orchestrating bacterial motility. All planctomycetes included in this study are motile in 454 one stage of their life cycle, while I. pallida lacks a flagellum and its motility is achieved 455 through gliding (16). Thus, chemotaxis is to be expected for all analyzed bacteria. We 456 focused on the distribution and abundance of two central chemotaxis proteins, so-called 457 methyl-accepting chemoreceptor proteins (MCPs) and the CheA-like HK, in order to gain 458 some insight into the motility intelligence between different species. While the first gives 459 an indication on the diversity of different stimuli sensed, the second is a reliable measure 460 for the number of independent chemotaxis systems encoded in a given bacterial genome. 461 Two major conclusions can be drawn from the data shown in Table 3. First, there is a 462 huge variance in the complexity of the chemotaxis systems between the eight species, 463 which is mostly independent of genome size. While B. marina, P. staleyi, and especially 464 G. obscuriglobus have a high “motility intelligence” with four to five independent 465 chemotaxis systems and 20-60 MCPs, the remaining planctomycetes only contain a very 466 limited number of chemotaxis proteins, based on one or two chemotaxis modules. Second, 467 our analysis revealed the complete lack of any chemotaxis machinery in the motile and 468 flagellated organism R. baltica, which is in agreement with previous studies (17). This 469 can mean that R. baltica can swim, but has no means to direct its movement, or it 470 somehow has acquired, developed or adapted a completely different mechanism for 471 adjusting its flagella motor. Both alternatives seem equally unlikely but point towards 472 very interesting future experiments in this planctomycetal model organism. 473 Extracytoplasmic function σ factors (ECFs). 474 ECFs belong to the pool of alternative σ factors that recognize specific promoter 475 sequences to control gene expression. Their activity is usually regulated by cognate anti-476 σ factors (5, 19). Recently, a classification of ECFs was published, based on about 400 477


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

19

microbial genomes (53). Giving the genome sequencing bias towards model bacteria, or 478 those with medical or biotechnological relevance, many phyla were underrepresented, 479 despite their ecological relevance. Not surprisingly, most ECFs from the dominant phyla 480 can be classified, while a majority of ECFs from underrepresented phyla, including the 481 Planctomycetes, could not be assigned to any defined ECF group (53). Because of the 482 high abundance of unclassified ECFs in Planctomycetes, we performed an in-depth 483 analysis of these proteins mimicking the initial ECF classification (53). The overall 484 abundance of ECFs in the eight sequenced planctomycetal species is shown in Fig. 7A. 485 Branches of the phylogenetic tree corresponding to known or novel conserved ECF 486 groups are color-coded in Fig. 7B and their genomic abundance and distribution is 487 indicated by the same colors in Fig. 7A. ECFs were found to represent a widely 488 distributed signaling principle in the phylum Planctomycetes. G. obscuriglobus in 489 particular is one of the ECF-richest bacteria sequenced to date, with its genome encoding 490 115 ECFs. This is only three proteins short of the current record holder, the delta-491 proteobacterium Plesiocystis pacifica. Based on the original set of 43 group-specific 492 HMMs described by Staroń et al. (2009), only 96 out of the 362 proteins could be 493 classified to one of four conserved ECF groups present in planctomycetes (data not 494 shown). The majority of these belong to the highly diverse group ECF01 of RpoE-like σ 495 factors that are functionally linked to typical membrane-anchored anti-σ factors 496 harboring an anti-sigma domain. The other classifiable ECFs were either assigned to the 497 two poorly characterized groups ECF22 and ECF42, which both lack a designated anti-σ 498 factor, or to the more distantly related ECF-like proteins of group ECF43, which seem to 499 be functionally linked to STPKs (53). In this new planctomycetal-specific analysis, we 500 defined eight novel ECF groups based on phylogenetic relationship and genomic context 501 conservation, raising the count of classified planctomycetal ECFs by a factor of three 502 from 96 to 304 (Tab. S5, Fig. 7, see supporting text for details). Most strikingly, the 503 branch ECF01-Gob (Fig. 7B light green) contains more than half of all ECFs encoded in 504 the genome of G. obscuriglobus (62 of 115) but not a single protein from any of the other 505 seven planctomycetal species. This strongly suggests that more than 50% of all ECFs in 506 this organism are evolutionary young paralogs. In addition, these Gemmata-specific 507 ECFs comprise a diverse and unusual domain architecture (Fig. 6B). They share very 508


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

20

large C-terminal extensions, sometimes exceeding the length of the ECF core by a factor 509 of more than four (19). Extension of about 20 ECF01-Gob proteins contain multiple 510 WD40 repeats that might be involved in the assembly of multi-protein complexes (51). In 511 addition, most ECF01-Gob proteins are membrane-anchored, harboring 1-3 putative 512 transmembrane helices between the ECF core and the C-terminal extension. These are the 513 first membrane-anchored ECFs described in bacteria. Giving the presence of a nucleus-514 like cellular organization in this organism, it is a highly attractive, but also purely 515 speculative hypothesis that this unusual type of transmembrane ECFs have specifically 516 evolved in this species to facilitate signal transduction between the cytoplasm and the 517 nucleoid. 518 Besides ECF01-Gob we identified four well-separated novel ECF groups (ECFSTK1 to -519 STK4) that are all functionally linked to STPKs, according to genomic context 520 conservation. Such a link was so far only described for the ECF-like proteins of group 521 ECF43 (53). This result demonstrates the wide distribution of signaling modules 522 consisting of a protein kinase and an ECF-type σ factor. While it is tempting to speculate 523 that in these modules, the STPKs function as sensors that phosphorylate and thereby 524 activate their cognate ECF partner, such a novel mechanism of ECF-dependent signal 525 transduction awaits experimental verification. 526 Taken together, our ECF analysis from the phylum Planctomycetes indicates that the 527 exploration of this third pillar of bacterial signal transduction has only just begun. 528 Especially in phyla underrepresented in the data set of the initial analysis (53), there is a 529 large potential for finding novel conserved mechanisms not present in the more common 530 model organisms. The groups described above not only add to the growing list of 531 conserved ECFs, but also allow a first glimpse into how planctomycetes employ such 532 proteins as a central mechanism of signal transduction. 533 534 Conclusion 535 Employing various bioinformatic methods, we identified proteins putatively involved in 536 planctomycetal cell division and in their conspicuous cell plan. In addition, we identified 537


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

21

planctomycetal signal transduction systems as well as sigma factors and evidence 538 towards a new signaling logic among Planctomycetes. As genetic manipulation of 539 Planctomycetes remains tedious, our findings will help to guide future wet-lab 540 experiments with a manageable list of the most promising target proteins to study 541 planctomycetal hallmark traits. 542


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

22

References 543 1. Bauer, M., T. Lombardot, H. Teeling, N. L. Ward, R. I. Amann, and F. O. 544

Glöckner. 2004. Archaea-like genes for C1-transfer enzymes in Planctomycetes: 545 Phylogenetic implications of their unexpected presence in this phylum. J Mol 546 Evol 59:571-586. 547

548 2. Bernander, R., and T. J. Ettema. 2010. FtsZ-less cell division in Archaea and 549

Bacteria. Curr Opin Microbiol 13:747-752. 550 551 3. Bertin, A., M. A. McMurray, J. Pierson, L. Thai, K. L. McDonald, E. A. Zehr, 552

G. Garcia, 3rd, P. Peters, J. Thorner, and E. Nogales. 2012. Three-553 dimensional ultrastructure of the septin filament network in Saccharomyces 554 cerevisiae. Mol Biol Cell 23:423-432. 555

556 4. Brochier, C., and H. Philippe. 2002. Phylogeny: A non-hyperthermophilic 557

ancestor for bacteria. Nature 417:244. 558 559 5. Butcher, B. G., T. Mascher, and J. D. Helmann. 2008. Environmental sensing 560

and the role of extracytoplasmic function (ECF) sigma factors., p. 233-261. In W. 561 M. El-Sharoud (ed.), Bacterial Physiology - a Molecular Approach Springer-562 Verlag GmbH, Berlin, Heidelberg. 563

564 6. Cayrou, C., D. Raoult, and M. Drancourt. 2010. Broad-spectrum antibiotic 565

resistance of Planctomycetes organisms determined by Etest. J Antimicrob 566 Chemother 65:2119-2122. 567

568 7. de Been, M., C. Francke, R. J. Siezen, and T. Abee. 2011. Novel sigmaB 569

regulation modules of Gram-positive bacteria involve the use of complex hybrid 570 histidine kinases. Microbiology 157:3-12. 571

572 8. de Hoon, M. J., P. Eichenberger, and D. Vitkup. 2010. Hierarchical evolution 573

of the bacterial sporulation network. Curr Biol 20:R735-745. 574 575 9. Felsenstein, J. 1989. Mathematics vs. Evolution: Mathematical Evolutionary 576

Theory. Science 246:941-942. 577 578 10. Forterre, P., and S. Gribaldo. 2010. Bacteria with a eukaryotic touch: A glimpse 579

of ancient evolution? Proc Natl Acad Sci U S A 107:12739-12740. 580 581 11. Fuchsman, C. A., and G. Rocap. 2006. Whole-genome reciprocal BLAST 582

analysis reveals that planctomycetes do not share an unusually large number of 583 genes with Eukarya and Archaea. Appl Environ Microbiol 72:6841-6844. 584

585 12. Fuerst, J. A., and E. Sagulenko. 2011. Beyond the bacterium: Planctomycetes 586

challenge our concepts of microbial structure and function. Nature Reviews. 587 Microbiology 9:403-413. 588

589 13. Fuerst, J. A., and R. I. Webb. 1991. Membrane-bounded nucleoid in the 590

eubacterium Gemmata obscuriglobus. Proc Natl Acad Sci U S A 88:8184-8188. 591 592 14. Galperin, M. Y. 2010. Diversity of structure and function of response regulator 593

output domains. Curr Opin Microbiol 13:150-159. 594


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

23

15. Gao, R., and A. M. Stock. 2009. Biological insights from structures of two-595 component proteins. Annu Rev Microbiol 63:133-154. 596

597 16. Giovannoni, S. J., W. Godchaux, 3rd, E. Schabtach, and R. W. Castenholz. 598

1987. Cell wall and lipid composition of Isosphaera pallida, a budding 599 eubacterium from hot springs. J Bacteriol 169:2702-2707. 600

601 17. Glöckner, F. O., M. Kube, M. Bauer, H. Teeling, T. Lombardot, W. Ludwig, 602

D. Gade, A. Beck, K. Borzym, K. Heitmann, R. Rabus, H. Schlesner, R. 603 Amann, and R. Reinhardt. 2003. Complete genome sequence of the marine 604 planctomycete Pirellula sp. strain 1. Proc Natl Acad Sci U S A 100:8298-8303. 605

606 18. Hall, T. 1999. BioEdit: a user-friendly biological sequence alignment editor and 607

analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 608 41:95-98. 609

610 19. Helmann, J. D. 2002. The extracytoplasmic function (ECF) sigma factors. Adv 611

Microb Physiol 46:47-110. 612 613 20. Ingram, B. O., C. Sohlenkamp, O. Geiger, and C. R. Raetz. 2010. Altered lipid 614

A structures and polymyxin hypersensitivity of Rhizobium etli mutants lacking the 615 LpxE and LpxF phosphatases. Biochem Biophys Acta 1801:593-604. 616

617 21. Jermy, A. 2010. Evolution: Bacterial endocytosis uncovered. Nature Reviews 618

Microbiology 8:534. 619 620 22. Jogler, C., F. O. Glockner, and R. Kolter. 2011. Characterization of 621

Planctomyces limnophilus and development of genetic tools for its manipulation 622 establish it as a model species for the phylum Planctomycetes. Appl Environ 623 Microbiol 77:5826-5829. 624

625 23. Jun, S. R., G. E. Sims, G. A. Wu, and S. H. Kim. 2010. Whole-proteome 626

phylogeny of prokaryotes by feature frequency profiles: An alignment-free 627 method with optimal feature resolution. Proc Natl Acad Sci U S A 107:133-138. 628

629 24. Kerger, B. D., C. A. Mancuso, P. D. Nichols, D. C. White, T. Langworthy, M. 630

Sittig, H. Schlesner, and P. Hirsch. 1988. The budding bacteria, Pirellula and 631 Planctomyces, with atypical 16S ribosomal-RNA and absence of peptidoglycan, 632 show Eubacterial phospholipids and uniquely high proportions of long-chain beta-633 hydroxy fatty-acids in the lipopolysaccharide Lipid-A. Archives of Microbiology 634 149:255-260. 635

636 25. Kirkpatrick, C. L., and P. H. Viollier. 2012. Decoding Caulobacter 637

development. FEMS microbiology reviews 36:193-205. 638 639 26. König, E., H. Schlesner, and P. Hirsch. 1984. Cell-wall studies on budding 640

bacteria of the Planctomyces Pasteuria group and on a Prosthecomicrobium sp. 641 Archives of Microbiology 138:200-205. 642

643 27. Liesack, W., H. Konig, H. Schlesner, and P. Hirsch. 1986. Chemical-644

composition of the peptidoglycan-free cell envelopes of budding bacteria of the 645 Pirella Planctomyces group. Archives of Microbiology 145:361-366. 646


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

24

28. Lindsay, M. R., R. I. Webb, and J. A. Fuerst. 1997. Pirellulosomes: A new type 647 of membrane-bounded cell compartment in planctomycete bacteria of the genus 648 Pirellula. Microbiology-Uk 143:739-748. 649

650 29. Lonhienne, T. G., E. Sagulenko, R. I. Webb, K. C. Lee, J. Franke, D. P. 651

Devos, A. Nouwens, B. J. Carroll, and J. A. Fuerst. 2010. Endocytosis-like 652 protein uptake in the bacterium Gemmata obscuriglobus. Proc Natl Acad Sci U S 653 A 107:12883-12888. 654

655 30. Lopez, D., and R. Kolter. 2010. Extracellular signals that define distinct and 656

coexisting cell fates in Bacillus subtilis. FEMS Microbiol Rev 34:134-149. 657 658 31. Ludwig, W., O. Strunk, R. Westram, L. Richter, H. Meier, Yadhukumar, A. 659

Buchner, T. Lai, S. Steppi, G. Jobb, W. Forster, I. Brettske, S. Gerber, A. W. 660 Ginhart, O. Gross, S. Grumann, S. Hermann, R. Jost, A. Konig, T. Liss, R. 661 Lussmann, M. May, B. Nonhoff, B. Reichel, R. Strehlow, A. Stamatakis, N. 662 Stuckmann, A. Vilbig, M. Lenke, T. Ludwig, A. Bode, and K. H. Schleifer. 663 2004. ARB: A software environment for sequence data. Nucleic Acids Res 664 32:1363-1371. 665

666 32. Lutkenhaus, J. 2007. Assembly dynamics of the bacterial MinCDE system and 667

spatial regulation of the Z ring. Annu Rev Biochem 76:539-562. 668 669 33. Majdalani, N., and S. Gottesman. 2005. The Rcs phosphorelay: a complex 670

signal transduction system. Annu Rev Microbiol 59:379-405. 671 672 34. Makarova, K. S., and E. V. Koonin. 2010. Two new families of the FtsZ-tubulin 673

protein superfamily implicated in membrane remodeling in diverse bacteria and 674 archaea. Biol Direct 5:33. 675

676 35. Makarova, K. S., N. Yutin, S. D. Bell, and E. V. Koonin. 2010. Evolution of 677

diverse cell division and vesicle formation systems in Archaea. Nature Reviews 678 Microbiology 8:731-741. 679

680 36. Marchler-Bauer, A., S. Lu, J. B. Anderson, F. Chitsaz, M. K. Derbyshire, C. 681

DeWeese-Scott, J. H. Fong, L. Y. Geer, R. C. Geer, N. R. Gonzales, M. 682 Gwadz, D. I. Hurwitz, J. D. Jackson, Z. Ke, C. J. Lanczycki, F. Lu, G. H. 683 Marchler, M. Mullokandov, M. V. Omelchenko, C. L. Robertson, J. S. Song, 684 N. Thanki, R. A. Yamashita, D. Zhang, N. Zhang, C. Zheng, and S. H. 685 Bryant. 2011. CDD: A Conserved Domain Database for the functional annotation 686 of proteins. Nucleic Acids Res 39:D225-229. 687

688 37. Margolis, P., A. Driks, and R. Losick. 1991. Establishment of cell type by 689

compartmentalized activation of a transcription factor. Science (New York, N.Y 690 254:562-565. 691

692 38. McCormick, J. R., and K. Flardh. 2012. Signals and regulators that govern 693

Streptomyces development. FEMS Microbiol Rev 36:206-231. 694 695 39. Mingorance, J., J. Tamames, and M. Vicente. 2004. Genomic channeling in 696

bacterial cell division. J Mol Recognit 17:481-487. 697


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

25

40. Offner, S., G. Wanner, and F. Pfeifer. 1996. Functional studies of the 698 gvpACNO operon of Halobacterium salinarium reveal that the GvpC protein 699 shapes gas vesicles. J Bacteriol 178:2071-2078. 700

701 41. Park, K. T., W. Wu, K. P. Battaile, S. Lovell, T. Holyoak, and J. Lutkenhaus. 702

2011. The Min oscillator uses MinD-dependent conformational changes in MinE 703 to spatially regulate cytokinesis. Cell 146:396-407. 704

705 42. Pearson, A., M. Budin, and J. J. Brocks. 2003. Phylogenetic and biochemical 706

evidence for sterol synthesis in the bacterium Gemmata obscuriglobus. Proc Natl 707 Acad Sci U S A 100:15352-15357. 708

709 43. Pereira, G., and Y. M. Yamashita. 2011. Fly meets yeast: Checking the correct 710

orientation of cell division. Trends Cell Biol 21:526-533. 711 712 44. Pereira, S. F., L. Goss, and J. Dworkin. 2011. Eukaryote-like serine/threonine 713

kinases and phosphatases in bacteria. Microbiol Mol Biol Rev MMBR 75:192-714 212. 715

716 45. Pilhofer, M., M. S. Ladinsky, A. W. McDowall, G. Petroni, and G. J. Jensen. 717

2011. Microtubules in bacteria: Ancient tubulins build a five-protofilament 718 homolog of the eukaryotic cytoskeleton. PLoS Biol 9:e1001213. 719

720 46. Pilhofer, M., K. Rappl, C. Eckl, A. P. Bauer, W. Ludwig, K. H. Schleifer, and 721

G. Petroni. 2008. Characterization and evolution of cell division and cell wall 722 synthesis genes in the bacterial phyla Verrucomicrobia, Lentisphaerae, 723 Chlamydiae, and Planctomycetes and phylogenetic comparison with rRNA genes. 724 J Bacteriol 190:3192-3202. 725

726 47. Pruesse, E., C. Quast, K. Knittel, B. M. Fuchs, W. Ludwig, J. Peplies, and F. 727

O. Glockner. 2007. SILVA: A comprehensive online resource for quality 728 checked and aligned ribosomal RNA sequence data compatible with ARB. 729 Nucleic Acids Res 35:7188-7196. 730

731 48. Santarella-Mellwig, R., J. Franke, A. Jaedicke, M. Gorjanacz, U. Bauer, A. 732

Budd, I. W. Mattaj, and D. P. Devos. 2010. The compartmentalized bacteria of 733 the Planctomycetes-Verrucomicrobia-Chlamydiae superphylum have membrane 734 coat-like proteins. PLoS Biol 8:e1000281. 735

736 49. Shaevitz, J. W., and Z. Gitai. 2010. The structure and function of bacterial Actin 737

homologs. Cold Spring Harb Perspect Biol 2. 738 739 50. Shi, X., S. Wegener-Feldbrugge, S. Huntley, N. Hamann, R. Hedderich, and 740

L. Sogaard-Andersen. 2008. Bioinformatics and experimental analysis of 741 proteins of two-component systems in Myxococcus xanthus. J Bacteriol 190:613-742 624. 743

744 51. Smith, T. F., C. Gaitatzes, K. Saxena, and E. J. Neer. 1999. The WD repeat: A 745

common architecture for diverse functions. Trends Biochem Sci 24:181-185. 746 747 52. Staron, A., and T. Mascher. 2010. General stress response in alpha-748

proteobacteria: PhyR and beyond. Mol Microbiol 78:271-277. 749


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

26

53. Staron, A., H. J. Sofia, S. Dietrich, L. E. Ulrich, H. Liesegang, and T. 750 Mascher. 2009. The third pillar of bacterial signal transduction: Classification of 751 the extracytoplasmic function (ECF) sigma factor protein family. Mol Microbiol 752 74:557-581. 753

754 54. Tamames, J., M. Gonzalez-Moreno, J. Mingorance, A. Valencia, and M. 755

Vicente. 2001. Bringing gene order into bacterial shape. Trends Genet 17:124-756 126. 757

758 55. Teeling, H., T. Lombardot, M. Bauer, W. Ludwig, and F. O. Glockner. 2004. 759

Evaluation of the phylogenetic position of the planctomycete 'Rhodopirellula 760 baltica' SH 1 by means of concatenated ribosomal protein sequences, DNA-761 directed RNA polymerase subunit sequences and whole genome trees. Int J Syst 762 Evol Microbiol 54:791-801. 763

764 56. Ulrich, L. E., E. V. Koonin, and I. B. Zhulin. 2005. One-component systems 765

dominate signal transduction in prokaryotes. Trends Microbiol 13:52-56. 766 767 57. Ulrich, L. E., and I. B. Zhulin. 2010. The MiST2 database: A comprehensive 768

genomics resource on microbial signal transduction. Nucleic Acids Res 38:D401-769 407. 770

771 58. van Niftrik, L., W. J. Geerts, E. G. van Donselaar, B. M. Humbel, R. I. Webb, 772

H. R. Harhangi, H. J. Camp, J. A. Fuerst, A. J. Verkleij, M. S. Jetten, and M. 773 Strous. 2009. Cell division ring, a new cell division protein and vertical 774 inheritance of a bacterial organelle in Anammox Planctomycetes. Mol Microbiol 775 73:1009-1019. 776

777 59. Wagner, M., and M. Horn. 2006. The Planctomycetes, Verrucomicrobia, 778

Chlamydiae and sister phyla comprise a superphylum with biotechnological and 779 medical relevance. Curr Opin Biotechnol 17:241-249. 780

781 60. Wang, D., S. S. Korban, and Y. Zhao. 2009. The Rcs phosphorelay system is 782

essential for pathogenicity in Erwinia amylovora. Mol Plant Pathol 10:277-290. 783 784 61. Weisburg, W. G., T. P. Hatch, and C. R. Woese. 1986. Eubacterial origin of 785

Chlamydiae. J Bacteriol 167:570-574. 786 787 62. Woese, C. R. 1987. Bacterial evolution. Microbiol Rev 51:221-271. 788


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

27

Figure and table legends 789 Figure 1: The class Planctomycetia is split into two distinct orders. 790 Maximum likelihood phylogenetic tree of planctomycetal 16S rRNA gene sequences. 791 Bootstrap values are only shown for deep branches and different Escherichia coli 792 sequences were used as outgroup represented by an arrow (accession numbers of used E. 793 coli sequences: GU594315, GU594305, GU594306, GU594304, GU594302, GU594316). 794 795 Figure 2: The planctomycetal core-genome. 796 The planctomycetal core-genome was determined comparing all eight sequenced 797 Planctomycetes species against each other, using the BLAST algorithm with a coverage 798 >60% and 1e-5 e-value cutoff as parameters. The Venn-diagram visualizes the 564 799 clusters fulfilling these criteria. After in silico subtraction, using proteins encoded by 800 Escherichia coli (accession: U00096) and Bacillus subtilis (accession: AL009126), 450 801 planctomycetal core genome clusters were eliminated and 114 planctomycetes specific 802 clusters remained. Genes within these 114 clusters fulfill two criteria: They are conserved 803 among planctomycetes and absent in genomes of the two most intensively studied model 804 organisms E. coli and B. subtilis. 805 806 Figure 3: Selected putative P. limnophilus operons. 807 Putative P. limnophilus operons 1-5 (PO) are shown, that fulfill two criteria: i) they 808 contain one gene that was evaluated as putatively involved in cell division or 809 compartmentalization based on manual inspection of the P. limnophilus core genome and 810 ii) harbor at least one more core-genome member. Core genome related genes are shown 811 as red outlined arrows. Identifiers for all genes are given and selected putative domains 812 are highlighted (vWF: Willebrand factor; FHS: Fork head associated). 813 814 815


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

28

Figure 4: Content and synteny of dcw operon genes among planctomycetes. 816 The cell division- (gray) and peptidoglycan synthesis- (black) related genes arranged in 817 the dcw operons of E. coli and B. subtilis are shown in comparison to homolog genes 818 from planctomycetal genomes. Genes of unknown function are shown in white. Genes 819 with weak similarity to the E. coli query are labeled in gray (W: mraW, Z: mraZ; L: ftsL). 820 821 Figure 5: Putative cell division related planctomycetal operons. 822 The putative operons, Plabr and Isop fulfilled two criteria: i) They contain a cell division- 823 or peptidoglycan synthesis related dcw gene and ii) at least one additional core-genome 824 member is present within the same putative operon. In a second round, corresponding P. 825 limnophilus genes were identified that turned out to localize on 5 different putative 826 operons (PO6-10). Selected protein domains are highlighted while core genome members 827 are boxed in red. 828 829 Figure 6: Domain architecture of special 2CSs and ECFs in planctomycetes. 830 The typical domain architecture is shown based on results from SMART database 831 (http://smart.embl-heidelberg.de/). (A) Novel type of RRs in Planctomycetes species 832 consisting of an N-terminal Ser/Thr/Tyr protein kinase domain (Pfam:Pkinase) and a C-833 terminal receiver domain. (B) Typical C-terminal extensions of group 01-Gob ECFs. 834 Standard ECFs (shown on the top) are comprised of only two regions σ2 835 (Pfam:Sigma70_r2) and σ4 (Pfam:Sigma70_r4) domains with a length of about 200 836 amino acids. The C-terminal extensions of group ECF01-Gob ECFs in planctomycetes 837 contain up to 1000 amino acids, which – amongst other domains (see text for detail) 838 contain putative transmembrane regions (shown in blue rectangle), as well as secretin and 839 WD40 domains with different arrangement style. Scale bar represents protein length in 840 amino acids. 841 842 843


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

29

Figure 7: Phylogenetic tree and classification of ECF sigma factors from planctomycetes. 844 Phylogenetic tree of all ECF sequences extracted from the MiST2 database generated by 845 the least squares distances method. A subsequent group analysis of unclassified 846 planctomycetal ECFs was done based on their phylogenetic distances, genomic context 847 conservation or presence of promoter motifs, which resulted in the assignment of the 848 novel groups ECF45, ECF46, ECFSTK1, ECFSTK2, ECFSTK3 and ECFSTK4, as well 849 as the ECF01-like groups, ECF01-Gob and ECF01-P. The corresponding sequence 850 information is collected in Tab. S5. Different groups are shown in different colors and 851 highlighted correspondingly in the tree map. (A) Abundance and distribution of ECFs in 852 the different planctomycetes species. Abbreviations: Rba, Rhodopirellula baltica SH 1; 853 Pst, Pirellula staleyi DSM 6068; Pma, Planctomyces maris DSM 8797; Pli, 854 Planctomyces limnophilus DSM 3776; Pbr, Planctomyces brasiliensis DSM 5305; Ipa, 855 Isosphaera pallida ATCC 43644; Gob, Gemmata obscuriglobus UQM 2246; Bma, 856 Blastopirellula marina DSM 3645. (B) Phylogenetic tree of ECFs in planctomycetes 857 species. Eight ECFs were excluded because of their phylogenetic distance to other ECFs. 858 These proteins are highlighted in Tab. S5. 859


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

30

Table 1: Selected features of planctomycetal genomes used in this study. 860 Selected features of all sequenced planctomycetal genomes used to determine the core-861 genome (CDS: coding sequence; PCG: protein coding genes; N.D. not determined). 862 863 Table 2: Cell division and cytoskeleton related genes of planctomycetes. 864 This table summarizes Tab. S4 while only those proteins are shown that are encoded in at 865 least one planctomycetal genome. Proteins are grouped by function while bold characters 866 describe dcw operon related proteins. 867 868 Table 3: Overview on planctomycetal signal transduction and regulation. 869 Overview/statistics on signal transducing and regulatory proteins encoded in the eight 870 planctomycetal genomes used throughout this study. Abbreviations: HK, histidine kinase; 871 HHK, hybrid HK; RR, response regulator; HRR, hybrid RR; STPK, Ser/Thr protein 872 kinase; TM, transmembrane. Numbers in parentheses indicate distribution of 873 planctomycete-specific RR illustrated in Fig. 7. The total number of chemotaxis proteins 874 does not include CheY-like RR, which cannot be distinguished from standard RR based 875 on their domain architecture. They are therefore included in RR numbers of 2CS. 876 on A

pril 4, 2019 by guesthttp://jb.asm

.org/D

ownloaded from

http://jb.asm.org/


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

ATPase domain

vWF

PO1

Plim_3217 Plim_3218Plim_3219

ATPase domain

PDZ

FHA domain

CARDB

PO1


Plim_3312

Plim_3311ç

Planctomycetal core genome

SpoIIE domain

CARDB

Trypsin domain

Plim_0821 Plim_0823Pli 0824 Pli 0825

PO2

unknown functionPlim_0820 Plim_0822 Plim_0824 Plim_0825

Pli 1420Pli 1411 Pli 1413

PO3

Plim_1409Plim_1420Plim_1411

Plim_1410 Plim_1412 Plim_1414Plim_1413

Plim_1415 Plim_1416 Plim_1418 Plim_1419Plim_1417

PO4


Plim_1803 Plim_1804 Plim_1805 Plim_1806Plim_1807


Plim_1811Plim_1801

ç ççPO5

Fig. 3


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

IpxCftsZftsAftsQddlmurCmurGftsWmurDmraYmurFmurEftsILWZ

E. coli

P. maris

B. subtilisZ W L murE mraY murD murG ftsZ ddl murFftsW murCftsI ftsW ftsQ

P. brasiliensis

murG ftsW mraY murF murE

murCmurGftsWmraYmurEmurF

ftsI

ftsI

W

W

Z

Z

D murC ddl

ddlmurD

P. limnophilusZ W ftsI ftsI murEmurF ftsW mraY murD murG ddl murC

P. staleyi

R. baltica

W murE mraY ddl

mraYW murE ddl

G. obscuriglobus

B. marinaWZ murE

I. pallida

ddl

W ddl

mraY W

mraY Fig. 4


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

PO 6

Plim_2791 Plim_2793 Plim_2794 Plim_2795

Plabr 0300ATP d i

Planctomycetal core genome

unknown function

Plabr FtsI

Plabr_0294_

Plabr_0301Plabr_0295

mraW

Plabr_0296 Plabr_0299 ATPase domain

DUF1501

Planctomycetes specific domains

PO7

Plim_0775 Plim_0777Plim_0776 Plim_0779Plim 0778 Plim 0780

Plim_0781

DUF1501

PSCyt1, pfam 07635

DUF1553, PSD1 Plim_0778 Plim_0780

Plim 3723 Plim 3725Plim_3724 Plim_3726

Plim 3727 Plim 0213 Plim 0214Plim_0215 Plim 0216

Plim_0212

DUF1549, PSCyt2

DUF1559, SBP_bac_10

Plim_3723 Plim_3725 Plim_3727 Plim_0213 Plim_0214 5 Plim_0216

PO8 PO9

Isop

Isop_3026 Isop_3036 Isop_3037

mraY-like

Isop_3038

Plim_26 Plim_33 Plim_34 Plim_35 Plim_36Plim_28 Plim_29Plim_31

Plim_32Plim_27 Plim_30

PO10

Fig. 5


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

(A)

(B)

Fig. 6g


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Table 1: Selected features of sequenced planctomycetal genomes except Anammox bacteria (PCG: protein coding genes; N.D. not determined)

Rhodopirellula baltica SH1

Planctomyces limnophilus

Pirellula staleyi

Isosphaera pallida

Planctomyces brasiliensis

Gemmata obscuriglobus

Planctomyces maris

Blastopirellula marina

Chromosome 7.1 MB 5.4 MB 6.2 MB 5.5 MB 6.0 MB 9.1 MB 7.8 MB 6.7 MB Sequence complete complete complete complete draft draft draft draft Accession NC_005027.1 NC_014148.1 NC_013720.1 NC_014962.1 NZ_AEIC00000000*) NZ_AEIC00000000 NZ_ABCE00000000 NZ_AANZ00000000 Plasmidsize - 37 kb - 56 kb - N.D. N.D. N.D. CDS 7403 4313 (+60) 4825 3791 (+32) 4865 8080 6530 6079 Genes 7325 4198 (+60) 4717 3690 (+32) 4769 7989 6480 6025 G + C contend

55.4 % 53.7 (57.0) % 57.5 % 62.4 (67.0) % 56 % 67.2 % 50.5 % 57.0 %

PCGs with function

34.22 % N.D. N.D. 59.33 % 57.55 % 39.48 % 40.66 % 39.64 % tRNAs 76 61 46 48 45 85 48 48 rRNA genes 3 4 3 9 6 6 1 6 References [1] [2] [3] [4] JGI, NCBI JGI, NCBI JGI, NCBI JGI, NCBI


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

1. Glöckner, F.O., et al., Complete genome sequence of the marine planctomycete Pirellula sp. strain 1. PNAS, U S A, 2003. 100(14): p. 8298-303. 2. Labutti, K., et al., Complete genome sequence of Planctomyces limnophilus type strain (Mu 290). Stand Genomic Sci, 2010. 3(1): p. 47-56. 3. Clum, A., et al., Complete genome sequence of Pirellula staleyi type strain (ATCC 27377). Stand Genomic Sci, 2009. 1(3): p. 308-16. 4. Göker, M., et al., Complete genome sequence of Isosphaera pallida type strain (IS1B). Stand Genomic Sci, 2011. 4(1): p. 63-71. *) downloaded from TrEMBL


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Table 2: Cell division and cytoskeleton related genes of planctomycetes.

This table summarizes Tab. S4 while only those proteins of are shown that are encoded in at least one planctomycetal genome. Proteins are grouped by function while bold characters describe dcw operon related proteins.

Rhod

opir

ellu

la

balti

ca S

H1

Plan

ctom

yces

lim

noph

ilus

Pire

llula

stal

eyi

Isos

phae

ra

palli

da

Plan

ctom

yces

br

asili

ensi

s

Gem

mat

a ob

scur

iglo

bus

Plan

ctom

yces

m

aris

Blas

topi

rellu

la

mar

ina

ClpX + + + + + + + + ClpP + + + + + + + + CpaE + + + + + + + + ddl + + + + + + + - EnvA + + + + + + + + FtsE + + + + + + + + FtsI - + - - + - + - FtsK + + + + + + + + FtsW - + - - + - + - RodA - + - - + - + - MraW + + + + + + + + MraY + + + + + + + + MraZ - + - - + - + + MreB - + - - + - + + MurC - + - - + - + - MurD - + - - + - + - MurE + + + - + - + + MurF - + - - + - + - MurG - + - - + - + - ParA + + + + + + + +


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

!"#$%&'( )*)$+ ,-./' 0/ 00/ 112 011 )*)$+3 45. 5678 )*)$+ -49:5;'!"#$%&'()*""+"#,-#)(.# /0/1 234 567 48 71 5/ 17,95: 3 74 6 4 46 8;*--#%#,&<$=+)(>"&<+$ 605/ /21 7?? 583 45 7/ ?2,94: 51 /8 76 1 551 /2@$&$'A#*)#,'#""(B# 1012 778 527 /5 51 54 25,95: 2 3 7 7 54 8C()*""+"#,$%#"*D( /07 7?1 512 27 77 / 41,95: 7 78 4 4 22 8C"#.=%&-D=*$,<)#$("(*.$($ /085 736 511 28 74 54 47,95: 1 1 5 5 7? 8C"#.=%&-D=*$,"(-.&'A("+$ 104/ 727 527 7? 78 ? 24,95: 4 / 7 5 72 8C"#.=%&-D=*$,-#)($ 303? 23/ 778 47 77 5? 43,95: 2 ? 5 5 17 8EA&B&'()*""+"#,<#"%(=# 3051 244 568 11 76 55 1/,97: / 7 8 8 4? 7

F"",.+-<*)$,G*)*,*H%)#=%*B,I)&-,%A*,J(KL7,B#%#<#$*F<<)*M(#%(&.$N,OP0,A($%(B(.*,Q(.#$*R,OOP0,AD<)(B,OPR,EE0,)*$'&.$*,)*>+"#%&)R,OEE0,AD<)(B,EER,KLCP0,K*)SLA),')&%*(.,Q(.#$*R,LJ0,%)#.$-*-<)#.*5N,.+-<*)$,(.,'#)*.%A*$*$,(.B(=#%*,B($%)(<+%(&.,&I,C"#.=%&-D=*%*T$'*=(I(=,EE,(""+$%)#%*B,(.,U(>V,W5F7N,LA*,%&%#",.+-<*),&I,=A*-&%#H($,')&%*(.,B&*$,.&%,(.="+B*,XA*YT"(Q*,EE0,GA(=A,=#..&%,<*,B($%(.>+($A*B,I)&-,$%#.B#)B,EE,<#$*B,&.,%A*(),B&-#(.,#)=A(%*=%+)*V,LA*D,#)*,%A*)*I&)*,(.="+B*B,(.,EE,.+-<*)$,&I,7XK

:5;'

-$<+7=3N,ZM*)M(*GSK%#%($%(=$,&.,$(>.#",%)#.$B+=(.>,#.B,)*>+"#%&)D,')&%*(.$,*.=&B*B,(.,C"#.=%&-D=*%*$,>*.&-*$

35,' 567(*)$?&'#7%*(7='&@7

25,'8AA


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

planctomycetes comparative genomics - american society for

Documents