Degenerative Evolution and Functional Diversification of Type-III Secretion Systems in the Insect Endosymbiont Sodalis glossinidius
http://www.100md.com
《分子生物学进展》
Department of Biology, University of Utah, Salt Lake City
Correspondence: E-mail: dale@biology.utah.edu.
Abstract
Sodalis glossinidius, a maternally transmitted endosymbiont of tsetse flies, maintains two phylogenetically distinct type-III secretion systems encoded by chromosomal symbiosis regions designated SSR-1 and SSR-2. Although both symbiosis regions are closely related to extant pathogenicity islands with similar gene inventories, SSR-2 has undergone novel degenerative adaptations in the transition to mutualism. Notably, SSR-2 lacks homologs of genes found in SSR-1 that encode secreted effector proteins known to facilitate the host cell cytoskeletal rearrangements necessary for bacterial entry and uptake into eukaryotic cells. Also, as a result of relaxed selection, SSR-2 has undergone inactivation of genes encoding components of the type-III secretion system needle substructure. In the current study, we used quantitative PCR to determine the expression profiles of ysaV (SSR-1) and invA (SSR-2) transcripts when S. glossinidius infects an insect cell line, and we used an invasion assay to characterize the phenotype of an S. glossinidius mutant that lacks the ability to produce an OrgA protein that is required for function of the SSR-2 secretome. Whereas SSR-1 is required for bacterial invasion of host cells and ysaV is expressed when bacteria contact host cells, SSR-2 is required for bacterial proliferation after entry, and invA is only expressed in the intracellular stage of infection. These results demonstrate that degenerative genetic adaptations in SSR-2 have promoted functional diversification of the Sodalis SSR-2 type-III secretion system.
Key Words: adaptive evolution ? type-III secretion system ? symbiosis ? neofunctionalization
Introduction
Many gram-negative bacterial pathogens and symbionts maintain intracellular associations with plant and animal hosts by utilizing specialized type-III protein secretion systems to deliver effector proteins into host cells (Hueck 1998). When deployed on the surface of the bacterial cell, the multicomponent type-III secretion system (TTSS) functions as a molecular syringe, enabling bacteria to inject effector proteins into host cells to facilitate their own entry (Collazo and Galan 1997). Although there is considerable diversity in the function of effector proteins secreted by different pathogens, their general role is to produce a more favorable environment for bacterial internalization (Cornelis and Van Gijsegem 2000).
The genetic components of type-III secretion systems (TTSSs) are organized on cognate gene clusters (termed "islands") found only in bacteria that have an intracellular lifestyle component. Phylogenetic analyses indicate that the TTSS genes share a common ancestor with the flagellar genes (Saier 2004). Despite the fact that all TTSSs share this common origin, there is considerable variation in the size, organization, and content of TTSS gene clusters in different bacterial lineages. Presumably this variation reflects functional adaptations occurring during coevolution between intracellular bacteria and their hosts.
At the present time, much of our understanding of type-III secretion is derived from the study of pathogenic bacteria and their interactions with plant and animal hosts. However, type-III secretion systems are also utilized by symbiotic bacteria participating in mutualistic associations with eukaryotic hosts (Marie, Broughton, and Deakin 2001; Dale et al. 2001, 2002; Horn et al. 2004). Although the function of the TTSS as a generalized protein translocation apparatus is likely to be conserved in pathogens and mutualists, the precise role of any given TTSS is defined by the effector proteins that are translocated to host cells.
To further our understanding of the role of type-III secretion in animal symbionts, we have focused on the insect endosymbiont Sodalis glossinidius. S. glossinidius is an intracellular symbiont of tsetse flies, relying on a predominantly vertical transmission strategy to facilitate its relationship with the tsetse host (Aksoy, Chen, and Hypsa 1997). In a previous study, it was shown that S. glossinidius utilizes a TTSS to coordinate invasion of host insect cells (Dale et al. 2001). The role of the TTSS appears to be critical in the symbiosis between S. glossinidius and the insect host because mutant S. glossinidius lacking the type-III secretion capability are unable to maintain a symbiotic association. Furthermore, molecular evolutionary data indicate that the acquisition of TTSS-endocing genes by S. glossinidius predated the establishment of a symbiotic relationship between S. glossinidius and its insect host (Dale et al. 2002).
In the current study, we obtained the complete nucleotide sequences of two distinct TTSS-encoding symbiosis regions in S. glossinidius, designated SSR-1 and SSR-2. Although SSR-1 and SSR-2 share a substantial number of homologous TTSS-encoding genes, genetic and molecular evolutionary analyses indicate that each region has a distinct ancestry and function. SSR-1 is most closely related to the ysa pathogenicity island found in Yersinia enterocolitica and maintains a full complement of TTSS-encoding genes, including genes predicted to encode effector proteins that facilitate host cell entry. SSR-2 is most closely related to the SPI-1 pathogenicity island found in Salmonella enterica but has a reduced gene inventory lacking any genes encoding effector proteins. Genes encoding essential protein components of the needle substructure have also been inactivated in SSR-2, indicating adaptation to a "needleless" export apparatus. Comparative gene expression assays indicate that the ysaV gene, located within SSR-1, is expressed when symbionts contact host cells, whereas invA, located within SSR-2, is expressed only when symbionts have entered host cells. In addition, a mutant S. glossinidius strain, lacking a functional orgA homolog within SSR-2, retains the ability to invade insect cells but is deficient in its ability to replicate in these cells after entry. These data are consistent with a role in host cell entry for SSR-1 and a role in postinvasion intracellular protein secretion for SSR-2.
Materials and Methods
Symbiont Culture, DNA Purification, and BAC Library Construction
S. glossinidius was grown in liquid culture, as described previously (Dale and Maudlin 1999). Log-phase cells were harvested by centrifugation (4,000 x g, 10 min) and DNA was prepared in agarose plugs using a CHEF bacterial DNA isolation kit (BioRad, Hercules, Calif.), according to the manufacturer's recommendations. DNA was partially digested with the restriction enzyme HindIII and the 90-kbp to 120-kbp digest fraction was recovered after pulsed-field gel electrophoresis. The S. glossinidius BAC library was constructed by Amplicon Express (Pullman, Wash.) using the pECBAC vector.
BAC Library Screening
Before screening, the S. glossinidius BAC library clones were inoculated into 100 μl aliquots of LB medium containing 12.5 μg/ml chloramphenicol and grown overnight in 96-well plates at 37°C. Templates were prepared for PCR screening by heat denaturation; 0.1 μl of medium from each overnight culture was inoculated into 20 μl of water and heated to 95°C for 10 min. PCR screening was performed with oligonucleotide primers known to amplify the TTSS genes invA and spaP (invAF; 5'-GAT AGG CGA TAA TCT GGT CGT-3', invAR1; 5'-AGG TGG GTG TAA ACT GTA AGC-3', and spaPF; 5'-CTG GAA AAC AGC ATG GAG TCC TAC-3', spaPR; 5'-ATA AAA GCC GAT TCT GAA GGA GTC-3'). PCR reactions were performed by combining each 20 μl aliquot of BAC clone template DNA with 20 μl of a 2 x PCR cocktail containing 5 mM MgCl2, 10 pmol of each primer, 0.4 mM dNTPs, and 2 units of Taq DNA polymerase. The PCR cycling conditions consisted of an initial denaturation step (2 min, 94°C) followed by 35 cycles of denaturation (20 s, 94°C), annealing (20 s, 58°C for invAF/R or 61°C for spaPF/R), and extension (30 s, 72°C) and a final extension (4 min, 72°C). PCRs were performed in 96-well plates until a total of eight 96-well plates had been screened, representing an approximately 32-fold coverage of the S. glossinidius genome size, based on an estimated genome size of 2.2 Mbp (Akman et al. 2001).
Shotgun Sequencing, Assembly, and Annotation
For large-scale DNA preparation, BAC clones were isolated and cultured in LB medium containing 12.5 μg/ml chloramphenicol. BAC DNA was prepared from 500-ml bacterial cultures using the Qiagen Large Construct Kit (Qiagen, Valencia, Calif.) according to the manufacturer's recommendations. Before shotgun library construction, BAC DNA was partially digested with the frequent cutting restriction enzyme CviJI (Gingrich, Boehrer, and Basu 1996). Blunt-ended DNA fragments in the 1-kbp to 2-kbp size range were purified by preparative electrophoresis, dephosphorylated with shrimp alkaline phosphatase, and cloned into the pCR-Blunt II-Topo vector (Invitrogen, Carlsbad, Calif.) according to the manufacturer's recommendations. For plasmid sequencing, template DNA was generated from recombinant clones by multiple displacement amplification (MDA) using the TempliPhi system (Amersham, Piscataway, NJ). Plasmids were sequenced until eightfold coverage was obtained for each BAC clone. Sequences were trimmed and assembled using the SeqMan program in the Lasergene package (DNAStar, Madison, Wis.). Underrepresented regions were sequenced directly from BAC clone DNA (Kelley et al. 1999). Putative open reading frames (ORFs) were identified using the GeneMark program (Lukashin and Borodovsky 1998) and translating Blast searches. Pseudogenes located between putative ORFs were identified by nucleotide Blast searches.
Phylogenetics and Molecular Evolutionary Analyses
Phylogenetic analyses were conducted on the nucleotide sequences of genes homologous to invA, invC, and the concatenated sequences of spaP, spaQ, and spaR in S. glossinidius and other gram-negative bacteria. Nucleotide sequence alignments were generated in TRANALIGN, based on protein sequence alignments generated in Clustal. All phylogenetic analyses were conducted on data sets that excluded the third codon position in each alignment. Initially, phylogenetic analyses were performed on large data sets comprising all of the invA, invC, and spaQ homologs available in the GenBank database using distance and parsimony methods. To obtain better resolution, we used maximum-likelihood (ML) methods in PAUP* (Swofford 2000) to analyze reduced data sets. Appropriate models of DNA sequence evolution were first estimated for each data set using the MODELTEST program (Posada and Crandall 1998). These evolutionary models were then utilized in exhaustive ML searches to identify phylogenetic trees. ML bootstrap analyses were used to evaluate trees by heuristic search using the tree-bisection reconnection (TBR) branch-swapping algorithm.
The frequencies of synonymous and nonsynonymous substitutions (dS and dN, respectively) were estimated by the Kumar method implemented in the MEGA version 2 program (Kumar et al. 2001) using pairwise alignments of homologous nucleotide sequences from S. glossinidius and Sitophilus zeamais primary endosymbiont (SZPE). Models of selection at the codon level were evaluated using Z-tests (Nei and Kumar 2000), implemented in MEGA2. For Z-tests, the estimates of variance for dN and dS were obtained by sampling 500 bootstrap replicates.
Gene Expression Assays—RNA Isolation
For expression assays, cells from a log-phase culture of S. glossinidius were inoculated into 5 ml growing cultures of Aedes albopictus C6/36 cells at a multiplicity of infection 10. Cultures were maintained at 25°C and harvested for RNA isolation immediately after inoculation of S. glossinidius and at time intervals 4 hours, 8 hours, 24 hours, and 48 hours after inoculation. Beore harvesting, cultures were examined by differential interference contrast (DIC) microscopy to determine the progress of cell invasion and to ensure that the cell cultures were free of contamination. To facilitate the rapid recovery of insect cells and bacteria for RNA isolation, insect cells were detached from culture flasks using a cell scraper, and all cellular material was pelleted immediately at 12,000 x g for 10 min at 25°C. After aspiration of culture media, cell pellets were snap frozen and stored at –70°C before RNA isolation. RNA was isolated from cell pellets using the RNAqueous isolation kit (Ambion, Austin, Tex.), according to the manufacturer's instructions. Isolated RNA was analyzed using the Agilent 2100 nanoanalyzer (Agilent, Palo Alto, Calif.) to confirm the presence of both 16S and 18S rRNA, derived from bacterial and insect total RNA, respectively.
Gene Expression Assays—Quantitative PCR
DNA contamination was removed from the RNA samples by multiple DNase I treatments until no DNA could be detected in TaqMan quantitative PCR assays using primers and probe that detect the S. glossinidius rplB gene. cDNA was prepared from DNA-free RNA samples by reverse transcription in 100 μl reactions (containing 800 ng of RNA) using the random primer approach with TaqMan reverse transcription reagents (Applied Biosystems, Foster City, Calif.).
Primers and probes for the TaqMan flourogenic 5' nuclease assay were designed using Primer Express software (Applied Biosystems) based on the S. glossinidius sequences obtained in the current study (SSR-1, ysaV; SSR-2, and invA) and in a previous study (rplB [Dale et al. 2002]). The sequences and optimal concentrations of the primers and probes used in the TaqMan assays are shown in table 1.
Table 1 Sequences and Optimal Concentrations of TaqMan Primers and Probes Used in This Study
TaqMan assays were conducted in 50 μl reactions using TaqMan Universal PCR Master Mix (Applied Biosystems) in an Applied Biosystems SDS 7000 sequence detection system, under standard cycling conditions (50°C for 2 min, 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 s and annealing/extension at 60°C for 1 min), using a ROX internal standard. The optimum concentration of TaqMan primers and probes (listed in table 1) were determined for each reaction according to the Applied Biosystems SDS 7000 user's manual.
TaqMan assays were conducted in triplicate using each primer/probe combination with 40 ng of cDNA as template. An internal standard curve was generated for each primer/probe combination using serially diluted S. glossinidius DNA as template. Three negative controls were included for each TaqMan assay to assess the integrity of reactions. Relative quantities of transcripts were estimated by the standard curve method (Dale et al. 2002) and the method (Livak and Schmittgen 2001). The use of the method was validated for each primer/probe combination by plotting CT for each primer/probe combination (using rplB as internal control) against log dilution of template DNA. The absolute values of the slopes of all plots were below 0.06, indicating that all primer/probe sets have similar amplification efficiencies (Livak and Schmittgen 2001).
Construction and Characterization of an orgA Mutant
To generate a double-crossover orgA mutant, the complete 1,943-bp prgK-orgA-orgAB sequence was amplified by PCR using primers prgKF (5'-ATGATGATCGCCATGCTGAGC) and orgABR (5'-TAAGGTTCATTTAGCGGC) and cloned into the pGEM T-vector (Promega, Madison, Wis.). The resulting recombinant plasmid DNA was digested with BamHI to cleave the single restriction site at position 46 in the orgA sequence. The linear plasmid was ligated to a kanamycin resistance cassette (Pharmacia, Piscataway, NJ) to generate plasmid pKANorgA. Electrocompetent S. glossinidius were transformed with 20 ng pKANorgA according to methods described previously (Dale et al. 2001). Because the ColE1 origin of replication does not function in S. glossinidius, pKANorgA serves a suicide vector for introduction of the prgK-orgA::kanR-orgAB construct. After electroporation, bacteria were resuspended in 5 ml MM medium and left to recover overnight at 25°C. On the following day, transformed bacteria were subcultured into MM medium containing 20 μg/ml kanamycin to select for recombinants. After overnight incubation in selective media, recombinants were plated onto MM agar containing 20 μg/ml kanamycin and maintained at 25°C in a sealed gas jar under microaerophilic conditions (5% O2, 10% CO2, and 85% N2). Double-crossover mutants were identified by screening kanamycin resistant colonies for ampicillin sensitivity. Kanamycin resistant, ampicillin sensitive recombinants were screened by PCR using primers prgKF and orgABR to confirm the insertional inactivation of orgA.
Invasion Assays
We used time-lapse DIC microscopy to monitor invasion of the C6/36 cell line by wild-type and mutant S. glossinidius. Slide flask monolayer cultures of C6/36 cells were infected with S. glossinidius at low multiplicity of infection (10) and maintained at 25°C. Counts were performed to determine the number of symbionts in host insect cells at time intervals after infection. These counts were performed on at least 25 insect cells within each flask, and the experiment was performed in triplicate to ensure the integrity of data.
Results
BAC Library Screening
Using two sets of PCR primers based on S. glossinidius TTSS sequences obtained during a previous study (Dale et al. 2001), we screened a BAC library constructed from S. glossinidius DNA. The PCR primer sets, known to amplify the invA and spaP homologs from S. glossinidius, each amplified DNA from distinct BAC clones. Complete shotgun sequencing of invA-positive and spaP-positive BAC clones revealed the presence of two TTSS-encoding symbiosis regions (SSR-1 and SSR-2) in different regions of the S. glossinidius genome. The organization of genes within SSR-1 and SSR-2 is shown in figure 1, and the sequences have been deposited in GenBank under accession numbers AY508229 and AY508228.
FIG. 1.— Organization of TTSS genes within the ysa pathogenicity island of Y. enterocolitica, the SSR-1 and SSR-2 regions of S. glossinidius, and the SPI-1 island of S. enterica. Genes are color coded in accordance with their predicted functions based on Blast homology. Putative transcriptional units are highlighted with arrows, indicating both the length and orientation of transcripts. Gene organization is conserved between the Y. enterocolitica ysa island and the S. glossinidius SSR-1 region and differs markedly from the S. glossinidius SSR-2 region and S. enterica SPI-1 island.
S. glossinidius Symbiosis Region 1
The 28.5-kbp SSR-1 is flanked by the rplS and acpM genes on the S. glossinidius chromosome. The current annotation indicates the presence of 23 intact open reading frames and three pseudogenes (defined as regions sharing uncharacteristically low levels of sequence identity with known TTSS genes, having one or more frame-shifting deletions or in-frame stop codons). SSR-1 is defined as a genetic island because it has a base composition (48.9 mol% G+C) substantially lower than the S. glossinidius chromosome (54.9 mol% G+C). In terms of gene organization, SSR-1 most closely resembles the TTSS-encoding ysa pathogenicity island found in Yersinia enterocolitica (Foultier et al. 2002). Both SSR-1 and the ysa island share an almost identical genetic structure and composition, including conservation of major transcriptional units. Because of these similarities, the genes within SSR-1 were named in accordance with their respective ysa homologs. The boundaries of SSR-1 are defined by operons encoding chaperones and effector proteins: the sigD-sigE operon located proximal to rplS and the sycB-yspB-yspC-yspD-yspA operon located proximal to acpM. Based on the available gene inventory, SSR-1 has all the protein components necessary to produce a functional secretion apparatus, including those effector proteins (YspB and YspC) homologous to the Sip proteins of Salmonella enterica that are known to facilitate host cytoskeletal modifications associated with bacterial invasion (Kaniga, Trollinger, and Galan 1995).
S. glossinidius Symbiosis Region 2
The second S. glossinidius symbiosis region, SSR-2, is flanked by the cusC and patA genes on the S. glossinidius chromosome, some 6.5 kbp upstream of the fli operon, encoding flagellar components. The base composition of SSR-2 is 56.3 mol% G+C and is similar to that of the S. glossinidius chromosome (54.9 mol% G+C). Based on ORF identification, GeneMark analysis, and Blast, SSR-2 is predicted to contain 16 intact ORFs and five pseudogenes. Gene organization within SSR-2 resembles that of the Salmonella enterica SPI-1 TTSS island, albeit in a reduced form. Notably, the S. glossinidius SSR-2 has no genes homologous to the sip genes found in S. enterica. In addition, SSR-2 lacks genes homologous to the iacP, iagB, and invH genes that encode an acyl carrier protein, a muramidase, and a chaperone, respectively, in S. enterica. According to the gene inventory, it seems unlikely that secretomes derived from S. glossinidius SSR-2 and S. enterica SPI-1 have comparable function because SSR-2 lacks the sip genes encoding effector proteins that facilitate SPI-1–mediated cell invasion in S. enterica (Kaniga, Trollinger, and Galan 1995). In addition, SSR-2 has no functional homologs of either the S. enterica SPI-1 invE gene, known to positively regulate the secretion of Sip proteins (Kubori and Galan 2002), or the S. enterica SPI-1 invB gene, known to encode a protein chaperone of SipA (Bronstein, Miao, and Miller 2000). Furthermore, SSR-2 lacks functional homologs of those S. enterica SPI-1 genes involved in regulation (spaN [Collazo, Zierler, and Galan 1995; Kubori et al. 2000]) and assembly (prgI, prgJ [Klein, Fahlen, and Jones 2000; Sukhan et al. 2001]) of the TTSS needle substructure. Thus, although SSR-2 maintains all genes necessary to produce the intracellular and membrane-bound components of the TTSS syringe, it lacks functional homologs of genes necessary for the needle substructure.
Ancestry of SSR-1 and SSR-2
Initially, we used parsimony and distance methods to construct TTSS gene trees with homologous sequences derived from a wide range of gram-negative bacteria. Because bootstrap analyses provided little or no support for deep relationships in these trees, we focused on a smaller, well-supported clade from within the initial data set. Subsequent analyses were based on ML approaches, incorporating models of nucleotide substitution derived from hierarchical ML ratio tests (Posada and Crandall 1998). To determine the ancestry of SSR-1 and SSR-2, we constructed gene trees from homologs of the S. enterica invA, invC, and concatenated spaPQR sequences (fig. 2). Only the invA tree was supported by more than 50% of ML bootstrap resamples at every node. The invA homolog from S. glossinidius SSR-1 was placed in a clade supported by 100% of bootstrap resamples with the ysaV gene from Y. enterocolitica and an invA homolog from the SZPE (Dale et al. 2002). The S. glossinidius SSR-2 invA homolog was also placed in a well-supported clade alongside invA sequences from Chromobacterium violaceum and S. enterica. For nodes with more than 50% bootstrap support, the invC tree showed the same overall topology. The S. glossinidius SSR-1 invC homolog was placed in a well-supported clade with the SZPE invC sequence and the Y. enterocolitica ysaN sequence. The S. glossinidius SSR-2 invC homolog was placed in a well-supported clade with the Chromobacterium violaceum and S. enterica invC sequences. Although the spaPQR tree was not as well resolved as the invA and invC trees, the S. glossinidius SSR-1 spaPQR sequence was also placed in a well-supported clade with the Y. enterocolitica ysa sequence, whereas the SSR-2 spaPQR sequence was placed in a clade with the spaPQR sequence from the weevil endosymbiont SZPE. This result is distinct because the SZPE invA and invC sequences were both placed in well-supported clades, along with their respective SSR-1 homologs. During construction of the invA, invC, and spaPQR trees, we included sequences from the E. coli 0157:H7 eiv/epa chromosomal island (O-island #115 [Perna et al. 2001]). This island is notable because, like SSR-2 in S. glossinidius, it also lacks genes encoding the Sip effector proteins. However, the invA, invC, and spaPQR trees do not support a direct ancestry of the E. coli 0157: H7 O-island #115 island and S. glossinidius SSR-2 after the loss of genes encoding the Sip effectors. Rather, it seems more likely that the Sip-encoding genes were lost independently in the lineages leading to S. glossinidius and E. coli 0157. In summary, the results of the phylogenetic analyses lead to three important conclusions. First, sequences from SSR-1 are closely related to homologs found in the Y. enterocolitica ysa island, whereas sequences from SSR-2 are closely related to homologs found in the S. enterica and C. violaceum SPI-1 islands. Second, the inv and spa homologs cloned previously from the weevil endosymbiont SZPE (Dale et al. 2002) most likely originated from two distinct chromosomal regions, analogous to SSR-1 and SSR-2, respectively. Third, the loss of genes encoding Sip effector proteins most likely occurred independently in S. glossinidius and E. coli 0157:H7.
FIG. 2.— Maximum-likelihood trees based on invA, invC, and the concatenated sequences of spaPQR homologs from S. glossinidius, SZPE, and selected enteric pathogens. Trees were generated by an exhaustive ML search process and bootstrapped by heuristic ML search. Bootstrap support values are included for nodes with greater than 50 % support. Trees were rooted with the E. coli flagellar homologs of the TTSS genes.
Genic Components of SSR-1 and SSR-2 Evolve by Purifying Selection
The occurrence of multiple pseudogenes within SSR-2 prompted us to determine the mode of selection operating on certain genes within SSR-1 and SSR-2. To avoid problems associated with saturation in the pairwise estimation of substitution frequencies, we were restricted to focusing on those SSR-1 and SSR-2 gene sequences that have been determined for S. glossinidius and the closely related weevil endosymbiont SZPE (table 2). To determine the modes of selection on these representative SSR-1 and SSR-2 genes, we first used a two-tailed Z-test to evaluate the hypothesis that these genes are evolving neutrally (dN dS). After rejection of this hypothesis for all genes (data not shown), we used a one-tailed Z-test to evaluate the hypothesis that these genes are evolving by purifying selection (dN < dS). This hypothesis was accepted for all genes with strong support (all P values < 0.005), indicating that genes in both SSR-1 and SSR-2 are evolving by purifying selection. Although these data indicate that SSR-2 has retained function despite the loss of genes encoding effectors and protein components of the syringe, we do not exclude the possibility that positive (diversifying) selection is operating locally at some sites within the TTSS-encoding genes.
Table 2 Sequence Divergence Between S. glossinidius and SZPE
Differential Expression of SSR-1 and SSR-2
We used quantitative TaqMan PCR assays to determine the expression profiles of the ysaV gene (located within SSR-1) and the invA gene (located within SSR-2) in S. glossinidius through a time course of infection in Aedes albopictus C6/36 cells. The numbers of transcripts generated from the ysaV and invA genes were measured, along with the numbers of transcripts of a control gene, rplB. Because rplB encodes a ribosomal protein, it is anticipated to be expressed at a constant level in bacteria undertaking translation. In addition, rplB is anticipated to be expressed at a relatively high level because it has a high codon adaptation index in bacteria (Dale et al. 2002; Lithwick and Margalit 2003). The relative abundance of each gene transcript was determined by estimation from an internal standard curve (Dale et al. 2002) and by the method (Livak and Schmittgen 2001). Although both methods of analysis yielded similar results, data generated by the standard-curve method, presented in figure 3, are considered more robust because plots of CT versus log dilution of template DNA produced slightly different slopes for each primer/probe combination tested.
FIG. 3.— Quantitative RT-PCR analysis of gene expression in S. glossinidius at time intervals throughout the initial stages of infection of A. albopictus C6/36 cells. The numbers of transcripts from ysaV, invA, and fusA were measured using the rplB gene as an endogenous control. "Fold-difference" values were determined by comparing transcript numbers in each sample to the 4-hour control sample. Peak levels of the ysaV transcript were detected 24 hours postinfection, when symbionts are actively invading insect cells. Peak levels of the invA transcript were detected 48 hours postinfection, when symbionts were established within host cells.
The expression patterns of ysaV and invA varied with respect to the timing of expression throughout the course of cell invasion. The ysaV gene was maximally expressed 24 hours after infection of the A. albopictus cell line. At this timepoint we observed a greater than 10-fold increase in the expression of ysaV relative to the 4-hour sample. Based on observations conducted in this study and in a previous study (Dale et al. 2001), this timepoint coincides with the invasion of insect cells by S. glossinidius. After invasion, at 48 hours postinfection, we observed a reduction in the numbers of ysaV transcripts, coincident with the establishment of the intracellular stage of infection. In contrast, we observed very little increase in the numbers of invA transcripts at 24 hours after infection of the A. albopictus cell line, when symbionts are most actively engaged in host cell invasion. Instead, invA was maximally expressed at 48 hours postinfection, when bacteria are established within host cells. At this timepoint, we detected a greater than 10-fold increase in the expression of invA relative to the 4-hour sample.
Mutant S. glossinidius Lacking orgA Demonstrate Impaired Replication in Host Cells
To investigate the function of SSR-2 in the process of cell invasion by S. glossinidius, we generated an orgA double-crossover knockout mutant. In S. enterica, the orgA gene is essential for SPI-1–mediated invasion, and the OrgA protein is predicted to be an essential component of the SPI-1 TTSS (Klein, Fahlen, and Jones 2000). Although we identified orgA homologs in both SSR-1 and SSR-2, these genes share a relatively low level of nucleotide sequence homology; hence, the inactivation of the SSR-2 orgA homolog is not expected to affect the function of the SSR-1 TTSS in S. glossinidius. To determine the phenotype of the S. glossinidius orgA mutant, we performed invasion assays in A. albopictus C6/36 monolayer cultures. Both mutant and wild-type S. glossinidius were monitored for their ability to invade and persist in A. albopictus cells over a period of 64 hours after infection (fig. 4). During the first 32 hours after infection, when S. glossinidius is engaged in the invasion of insect cells, we observed no difference in the numbers of orgA mutant and wild-type bacteria invading insect cells. However, after establishment of the intracellular infection, the orgA mutant was substantially impaired in its ability to replicate inside cells, relative to the wild-type strain of S. glossinidius. Taken together with the expression data, these results suggest that SSR-2 has a role distinct from that of SSR-1, enhancing proliferation of S. glossinidius inside insect cells, after SSR-1–mediated entry. In addition, the presence of a clear phenotype in the orgA mutant indicates that mutations leading to the formation of prgI and prgJ pseudogenes, have not eliminated the expression of genes (including orgA) that are located downstream in the putative prgH-orgAb polycistron.
FIG. 4.— Invasion assay in the A. albopictus C6/36 cell line. We counted the numbers of orgA mutant and wild-type S. glossinidius in insect cells at 8-hour intervals throughout the course of infection of C6/36 cells. Although the orgA mutant remains invasive, it has an impaired ability to replicate inside host cells after invasion. Counts were obtained from at least 25 insect cells in three replicate experiments at each timepoint.
Discussion
In the current study, we obtained the complete nucleotide sequences of two chromosomal symbiosis regions, designated SSR-1 and SSR-2, from the mutualistic insect endosymbiont, S. glossinidius. Although SSR-1 and SSR-2 both encode genetic components of type-III secretion, each region has a distinct base composition, gene content, and gene organization, indicative of independent ancestry. Phylogenetic analyses indicate that SSR-1 is most closely related to the ysa pathogenicity island of Y. enterocolitica, whereas SSR-2 is most closely related to the SPI-1 pathogenicity islands found in S. enterica and C. violaceum (Ribeiro de Vasconcelos et al. 2003).
Although TTSS genes have been identified previously in Sodalis and in the closely related weevil endosymbiont SZPE (Dale et al. 2001, 2002), there was no prior indication of the presence of phylogenetically distinct TTSS genes in separate chromosomal regions. However, the phylogenetic analyses presented in the current study indicate that the inv genes obtained in these previous studies were derived from SSR-1, whereas the spa genes were derived from SSR-2. Because it was clearly demonstrated that acquisition of both the inv and spa homologs predates the divergence of a common ancestor of S. glossinidius and SZPE (Dale et al. 2002), we can now assume that SSR-1 and SSR-2 were both present in a presymbiotic ancestor of S. glossinidius and SZPE.
Several bacterial pathogens are known to maintain two or more distinct type-III secretion systems that function in different ways to modulate pathogenesis. For example, S. enterica maintains two TTSS-encoding pathogenicity islands, designated SPI-1 and SPI-2 (Ochman and Groisman 1996). Whereas SPI-1 is expressed upon contact with host cells and facilitates bacterial invasion (Zhou and Galan 2001), SPI-2 is expressed only after invasion and promotes replication of bacteria in host-enclosed vacuoles (Waterman and Holden 2003).
In S. glossinidius, several compelling lines of evidence indicate that SSR-1 has an important role in mediating the invasion of host insect cells. First, SSR-1 maintains all genes necessary to produce a type-III secretome, including those genes encoding secreted effector proteins that facilitate host cytoskeletal modifications associated with invasion. Second, there is a substantial increase in the number of ysaV transcripts in S. glossinidius cells immediately before the invasion of insect cells in vitro. Third, mutant S. glossinidius lacking the SSR-1 ysaN gene are deficient in their ability to invade insect cells in vitro and cannot establish symbiosis in vivo (Dale et al. 2001).
Based on the gene inventory, SSR-2 lacks functional homologs of the spaN, prgI, and prgJ genes necessary for the production of the TTSS needle substructure in S. enterica. Although open reading frames can be detected in the anticipated positions of spaN, prgI, and prgJ, these reading frames are truncated and share uncharacteristically low levels of sequence identity with functional homologs in the public database. This clearly indicates relaxed selection on those genes encoding protein components of the needle. Because our molecular evolutionary analyses indicate that other genes in SSR-2 are evolving under strong purifying selection, it seems that SSR-2 retains functionality despite the absence of intact spaN, prgI, and prgJ genes. Notably, SSR-2 also lacks any genes encoding the Sip effector proteins that are translocated to host cells during SPI-1–mediated invasion by S. enterica. Furthermore, SSR-2 lacks functional copies of the invE and invB genes that are known to facilitate the translocation of Sip proteins in S. enterica (Bronstein, Miao, and Miller 2000; Kubori and Galan 2002). Although the absence of genes encoding effector proteins is striking, the E. coli 0157:H7 pathogen also lacks genes encoding Sip effector proteins within the eiv/epa (O-island #115) (Perna et al. 2001), predicted to encode TTSS proteins. Based on the phylogenetic analyses presented in the current study, the loss of genes encoding the Sip effector proteins in S. glossinidius and E. coli 0157:H7 occurred as a result of independent evolutionary events. Although many pathogens are known to secrete type-III effectors encoded by genes located outside of pathogenicity islands (Cornelis and Van Gijsegem 2000; Waterman and Holden 2003; Chang et al. 2004), these effectors have yet to be identified in S. glossinidius.
Because of the close phylogenetic relationship between genes encoding components of the TTSS in SSR-1 and SSR-2, it is pertinent to consider the possibility of functional complementation between the two secretion systems. Because SSR-2 has a reduced gene inventory relative to SSR-1, one could envisage a scenario in which the missing or inactive genes in SSR-2 are complemented by the respective functional homologs in SSR-1. This observation would imply a process of subfunctionalization (Hughes 1994; Lynch et al. 2001) occurring in SSR-2 as part of a degenerative adaptation to life in the insect host. However, the results of the current study contradict such a hypothesis and suggest instead that SSR-2 has evolved a new and independent function through neofunctionalization (Walsh 2003). This finding is evident from the results of quantitative PCR assays indicating differential expression of transcripts derived from SSR-1 and SSR-2 during invasion. Also, the phenotypes of mutants lacking key components of SSR-1 and SSR-2 are distinct; the SSR-1 ysaN mutant is deficient in its ability to invade insect cells (Dale et al. 2001), and the SSR-2 orgA mutant lacks the ability to replicate inside host cells after invasion.
Although SSR-2 is clearly a descendent of the SPI-1–like pathogenicity islands found in S. enterica and related pathogens, the function of SSR-2 has been modulated in S. glossinidius such that it now more closely resembles the SPI-2 secretion system of S. enterica. Adaptation by gene duplication and neofunctionalization is predicted to represent an important source of innovation in evolution (Ohno 1970). Indeed, these processes have been implicated in the invention of type-III secretion and the origin of pathogenesis in gram-negative bacteria (Horn et al. 2004; Saier 2004). In the context of symbiosis, such innovations could drive important evolutionary transitions, including the transition from parasitism to mutualism (Dale et al. 2001).
Acknowledgements
The authors thank Serap Aksoy (Yale University) for providing a culture of S. glossinidius. We also thank Jon Seger and an anonymous reviewer for providing useful comments.
References
Akman, L., R. V. Rio, C. B. Beard, and S. Aksoy. 2001. Genome size determination and coding capacity of Sodalis glossinidius, an enteric symbiont of tsetse flies, as revealed by hybridization to Escherichia coli gene arrays. J. Bacteriol. 183:4517–4525.
Aksoy, S., X. Chen, and V. Hypsa. 1997. Phylogeny and potential transmission routes of midgut-associated endosymbionts of tsetse (Diptera:Glossinidae). Insect Mol. Biol. 6:183–190.
Bronstein, P. A., E. A. Miao, and S. I. Miller. 2000. InvB is a type III secretion chaperone specific for SspA. J. Bacteriol. 182:6638–6644.
Chang, J. H., A. K. Goel, S. R. Grant, and J. L. Dangl. 2004. Wake of the flood: ascribing functions to the wave of type III effector proteins of phytopathogenic bacteria. Curr. Opin. Microbiol. 7:11–18.
Collazo, C. M., and J. E. Galan. 1997. The invasion-associated type-III protein secretion system in Salmonella: a review. Gene 192:51–59.
Collazo, C. M., M. K. Zierler, and J. E. Galan. 1995. Functional analysis of the Salmonella typhimurium invasion genes invl and invJ and identification of a target of the protein secretion apparatus encoded in the inv locus. Mol. Microbiol. 15:25–38.
Cornelis, G. R., and F. Van Gijsegem. 2000. Assembly and function of type III secretory systems. Annu. Rev. Microbiol. 54:735–774.
Dale, C., and I. Maudlin. 1999. Sodalis gen. nov. and Sodalis glossinidius sp. nov., a microaerophilic secondary endosymbiont of the tsetse fly Glossina morsitans morsitans. Int. J. Syst. Bacteriol. 49:267–275.
Dale, C., G. R. Plague, B. Wang, H. Ochman, and N. A. Moran. 2002. Type III secretion systems and the evolution of mutualistic endosymbiosis. Proc. Natl. Acad. Sci. USA 99:12397–12402.
Dale, C., S. A. Young, D. T. Haydon, and S. C. Welburn. 2001. The insect endosymbiont Sodalis glossinidius utilizes a type III secretion system for cell invasion. Proc. Natl. Acad. Sci. USA. 98:1883–1888.
Foultier, B., P. Troisfontaines, S. Muller, F. R. Opperdoes, and G. R. Cornelis. 2002. Characterization of the ysa pathogenicity locus in the chromosome of Yersinia enterocolitica and phylogeny analysis of type III secretion systems. J. Mol. Evol. 55:37–51.
Gingrich, J. C., D. M. Boehrer, and S. B. Basu. 1996. Partial CviJI digestion as an alternative approach to generate cosmid sublibraries for large-scale sequencing projects. Biotechniques 21:99–104.
Horn, M., A. Ollingro, S. Schmitz-Esser et al. (13 co-authors). 2004. Illuminating the evolutionary history of Chlamydiae. Science 304:728–730.
Hueck, C. J. 1998. Type III protein secretion systems in bacterial pathogens of animals and plants. Microbiol. Mol. Biol. Rev. 62:379–433.
Hughes, A. L. 1994. The evolution of functionally novel proteins after gene duplication. Proc. R. Soc. Lond. B Biol. Sci. 256:119–124.
Kaniga, K., D. Trollinger, and J. E. Galan. 1995. Homologues of the Shigella invasions ipaB and ipaC are required for Salmonella typhimurium entry into cultured cells. J. Bacteriol. 177:3965–3971.
Kelley, J. M., C. E. Fields, M. B. Craven, D. Bocskai, U. Kim, S. D. Rounsley, and M. D. Adams. 1999. High throughput direct end sequencing of BAC clones. Nucleic Acids Res. 27:1539–1546.
Klein, J. R., T. F. Fahlen, and B. D. Jones. 2000. Transcriptional organization and function of invasion genes within Salmonella enterica serovar typhimurium pathogenicity island 1, including the prgH, prgI, prgJ, prgK, orgA, orgB, and orgC genes. Infect. Immun. 68:3368–3376.
Kubori, T., and J. E. Galan. 2002. Salmonella type III secretion-associated protein InvE controls translocation of effector proteins into host cells. J. Bacteriol. 184:4699–4708.
Kubori, T., A. Sukhan, S. I. Aizawa, and J. E. Galan. 2000. Molecular characterization and assembly of the needle complex of the Salmonella typhimurium type III protein secretion system. Proc. Natl. Acad. Sci. USA. 97:10225–10230.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244–1245.
Lithwick, G., and H. Margalit. 2003. Hierarchy of sequence-dependent features associated with prokaryotic translation. Genome Res. 13:2665–2673.
Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-delta delta C(T)) method. Methods 25:402–408.
Lukashin, A., and M. Borodovsky. 1998. GeneMark.hmm: new solutions for gene finding. Nucleic Acids Res. 26:1107–1115.
Lynch, M., M. O'Hely, B. Walsh, and A. Force. 2001. Distribution of pathogenicity islands in Salmonella spp. Genetics 159:1789–1804.
Marie, C., W. J. Broughton, and W. J. Deakin. 2001. Rhizobium type III secretion systems: legume charmers or alarmers? Curr. Opin. Plant Biol. 4:336–342.
Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University Press, New York
Ochman, H., and E. A. Groisman. 1996. Salmonella entry into host cells: the work in concert of type III secreted effector proteins. Infect. Immun. 64:5410–5412.
Ohno, S. 1970. Evolution by gene duplication. Springer, Heidelberg.
Perna, N.T., G. Plunkett, V. Burland et al. (28 co-authors). 2001. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 409:529–533.
Posada, J., and K. A. Crandall. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14:817–818.
Ribeiro de Vasconcelosa, A. T., D. F. de Almeidab, M. Hungriac et al. (112 co-authors). 2003. The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability. Proc. Natl. Acad. Sci. USA 100:11660–11665.
Saier, M. H. 2004. Evolution of bacterial type III protein secretion systems. Trends Microbiol. 12:113–115.
Sukhan, A., T. Kubori, J. Wilson, and J. E. Galan. 2001. Genetic analysis of assembly of the Salmonella enterica serovar typhimurium type III secretion-associated needle complex. J. Bacteriol. 183:1159–1167.
Swofford, D. L. 2000. PAUP*: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, Mass.
Walsh, B. 2003. Population-genetic models of the fates of duplicate genes. Genetica 118:279–294.
Waterman, S. R., and D. W. Holden. 2003. Wake of the flood: ascribing functions to the wave of type III effector proteins of phytopathogenic bacteria. Cell. Microbiol. 5:501–511.
Zhou, D., and J. E. Galan. 2001. Functions and effectors of the Salmonella pathogenicity island 2 type III secretion system. Microbes Infect. 3:1293–1298.(Colin Dale, Tait Jones an)
Correspondence: E-mail: dale@biology.utah.edu.
Abstract
Sodalis glossinidius, a maternally transmitted endosymbiont of tsetse flies, maintains two phylogenetically distinct type-III secretion systems encoded by chromosomal symbiosis regions designated SSR-1 and SSR-2. Although both symbiosis regions are closely related to extant pathogenicity islands with similar gene inventories, SSR-2 has undergone novel degenerative adaptations in the transition to mutualism. Notably, SSR-2 lacks homologs of genes found in SSR-1 that encode secreted effector proteins known to facilitate the host cell cytoskeletal rearrangements necessary for bacterial entry and uptake into eukaryotic cells. Also, as a result of relaxed selection, SSR-2 has undergone inactivation of genes encoding components of the type-III secretion system needle substructure. In the current study, we used quantitative PCR to determine the expression profiles of ysaV (SSR-1) and invA (SSR-2) transcripts when S. glossinidius infects an insect cell line, and we used an invasion assay to characterize the phenotype of an S. glossinidius mutant that lacks the ability to produce an OrgA protein that is required for function of the SSR-2 secretome. Whereas SSR-1 is required for bacterial invasion of host cells and ysaV is expressed when bacteria contact host cells, SSR-2 is required for bacterial proliferation after entry, and invA is only expressed in the intracellular stage of infection. These results demonstrate that degenerative genetic adaptations in SSR-2 have promoted functional diversification of the Sodalis SSR-2 type-III secretion system.
Key Words: adaptive evolution ? type-III secretion system ? symbiosis ? neofunctionalization
Introduction
Many gram-negative bacterial pathogens and symbionts maintain intracellular associations with plant and animal hosts by utilizing specialized type-III protein secretion systems to deliver effector proteins into host cells (Hueck 1998). When deployed on the surface of the bacterial cell, the multicomponent type-III secretion system (TTSS) functions as a molecular syringe, enabling bacteria to inject effector proteins into host cells to facilitate their own entry (Collazo and Galan 1997). Although there is considerable diversity in the function of effector proteins secreted by different pathogens, their general role is to produce a more favorable environment for bacterial internalization (Cornelis and Van Gijsegem 2000).
The genetic components of type-III secretion systems (TTSSs) are organized on cognate gene clusters (termed "islands") found only in bacteria that have an intracellular lifestyle component. Phylogenetic analyses indicate that the TTSS genes share a common ancestor with the flagellar genes (Saier 2004). Despite the fact that all TTSSs share this common origin, there is considerable variation in the size, organization, and content of TTSS gene clusters in different bacterial lineages. Presumably this variation reflects functional adaptations occurring during coevolution between intracellular bacteria and their hosts.
At the present time, much of our understanding of type-III secretion is derived from the study of pathogenic bacteria and their interactions with plant and animal hosts. However, type-III secretion systems are also utilized by symbiotic bacteria participating in mutualistic associations with eukaryotic hosts (Marie, Broughton, and Deakin 2001; Dale et al. 2001, 2002; Horn et al. 2004). Although the function of the TTSS as a generalized protein translocation apparatus is likely to be conserved in pathogens and mutualists, the precise role of any given TTSS is defined by the effector proteins that are translocated to host cells.
To further our understanding of the role of type-III secretion in animal symbionts, we have focused on the insect endosymbiont Sodalis glossinidius. S. glossinidius is an intracellular symbiont of tsetse flies, relying on a predominantly vertical transmission strategy to facilitate its relationship with the tsetse host (Aksoy, Chen, and Hypsa 1997). In a previous study, it was shown that S. glossinidius utilizes a TTSS to coordinate invasion of host insect cells (Dale et al. 2001). The role of the TTSS appears to be critical in the symbiosis between S. glossinidius and the insect host because mutant S. glossinidius lacking the type-III secretion capability are unable to maintain a symbiotic association. Furthermore, molecular evolutionary data indicate that the acquisition of TTSS-endocing genes by S. glossinidius predated the establishment of a symbiotic relationship between S. glossinidius and its insect host (Dale et al. 2002).
In the current study, we obtained the complete nucleotide sequences of two distinct TTSS-encoding symbiosis regions in S. glossinidius, designated SSR-1 and SSR-2. Although SSR-1 and SSR-2 share a substantial number of homologous TTSS-encoding genes, genetic and molecular evolutionary analyses indicate that each region has a distinct ancestry and function. SSR-1 is most closely related to the ysa pathogenicity island found in Yersinia enterocolitica and maintains a full complement of TTSS-encoding genes, including genes predicted to encode effector proteins that facilitate host cell entry. SSR-2 is most closely related to the SPI-1 pathogenicity island found in Salmonella enterica but has a reduced gene inventory lacking any genes encoding effector proteins. Genes encoding essential protein components of the needle substructure have also been inactivated in SSR-2, indicating adaptation to a "needleless" export apparatus. Comparative gene expression assays indicate that the ysaV gene, located within SSR-1, is expressed when symbionts contact host cells, whereas invA, located within SSR-2, is expressed only when symbionts have entered host cells. In addition, a mutant S. glossinidius strain, lacking a functional orgA homolog within SSR-2, retains the ability to invade insect cells but is deficient in its ability to replicate in these cells after entry. These data are consistent with a role in host cell entry for SSR-1 and a role in postinvasion intracellular protein secretion for SSR-2.
Materials and Methods
Symbiont Culture, DNA Purification, and BAC Library Construction
S. glossinidius was grown in liquid culture, as described previously (Dale and Maudlin 1999). Log-phase cells were harvested by centrifugation (4,000 x g, 10 min) and DNA was prepared in agarose plugs using a CHEF bacterial DNA isolation kit (BioRad, Hercules, Calif.), according to the manufacturer's recommendations. DNA was partially digested with the restriction enzyme HindIII and the 90-kbp to 120-kbp digest fraction was recovered after pulsed-field gel electrophoresis. The S. glossinidius BAC library was constructed by Amplicon Express (Pullman, Wash.) using the pECBAC vector.
BAC Library Screening
Before screening, the S. glossinidius BAC library clones were inoculated into 100 μl aliquots of LB medium containing 12.5 μg/ml chloramphenicol and grown overnight in 96-well plates at 37°C. Templates were prepared for PCR screening by heat denaturation; 0.1 μl of medium from each overnight culture was inoculated into 20 μl of water and heated to 95°C for 10 min. PCR screening was performed with oligonucleotide primers known to amplify the TTSS genes invA and spaP (invAF; 5'-GAT AGG CGA TAA TCT GGT CGT-3', invAR1; 5'-AGG TGG GTG TAA ACT GTA AGC-3', and spaPF; 5'-CTG GAA AAC AGC ATG GAG TCC TAC-3', spaPR; 5'-ATA AAA GCC GAT TCT GAA GGA GTC-3'). PCR reactions were performed by combining each 20 μl aliquot of BAC clone template DNA with 20 μl of a 2 x PCR cocktail containing 5 mM MgCl2, 10 pmol of each primer, 0.4 mM dNTPs, and 2 units of Taq DNA polymerase. The PCR cycling conditions consisted of an initial denaturation step (2 min, 94°C) followed by 35 cycles of denaturation (20 s, 94°C), annealing (20 s, 58°C for invAF/R or 61°C for spaPF/R), and extension (30 s, 72°C) and a final extension (4 min, 72°C). PCRs were performed in 96-well plates until a total of eight 96-well plates had been screened, representing an approximately 32-fold coverage of the S. glossinidius genome size, based on an estimated genome size of 2.2 Mbp (Akman et al. 2001).
Shotgun Sequencing, Assembly, and Annotation
For large-scale DNA preparation, BAC clones were isolated and cultured in LB medium containing 12.5 μg/ml chloramphenicol. BAC DNA was prepared from 500-ml bacterial cultures using the Qiagen Large Construct Kit (Qiagen, Valencia, Calif.) according to the manufacturer's recommendations. Before shotgun library construction, BAC DNA was partially digested with the frequent cutting restriction enzyme CviJI (Gingrich, Boehrer, and Basu 1996). Blunt-ended DNA fragments in the 1-kbp to 2-kbp size range were purified by preparative electrophoresis, dephosphorylated with shrimp alkaline phosphatase, and cloned into the pCR-Blunt II-Topo vector (Invitrogen, Carlsbad, Calif.) according to the manufacturer's recommendations. For plasmid sequencing, template DNA was generated from recombinant clones by multiple displacement amplification (MDA) using the TempliPhi system (Amersham, Piscataway, NJ). Plasmids were sequenced until eightfold coverage was obtained for each BAC clone. Sequences were trimmed and assembled using the SeqMan program in the Lasergene package (DNAStar, Madison, Wis.). Underrepresented regions were sequenced directly from BAC clone DNA (Kelley et al. 1999). Putative open reading frames (ORFs) were identified using the GeneMark program (Lukashin and Borodovsky 1998) and translating Blast searches. Pseudogenes located between putative ORFs were identified by nucleotide Blast searches.
Phylogenetics and Molecular Evolutionary Analyses
Phylogenetic analyses were conducted on the nucleotide sequences of genes homologous to invA, invC, and the concatenated sequences of spaP, spaQ, and spaR in S. glossinidius and other gram-negative bacteria. Nucleotide sequence alignments were generated in TRANALIGN, based on protein sequence alignments generated in Clustal. All phylogenetic analyses were conducted on data sets that excluded the third codon position in each alignment. Initially, phylogenetic analyses were performed on large data sets comprising all of the invA, invC, and spaQ homologs available in the GenBank database using distance and parsimony methods. To obtain better resolution, we used maximum-likelihood (ML) methods in PAUP* (Swofford 2000) to analyze reduced data sets. Appropriate models of DNA sequence evolution were first estimated for each data set using the MODELTEST program (Posada and Crandall 1998). These evolutionary models were then utilized in exhaustive ML searches to identify phylogenetic trees. ML bootstrap analyses were used to evaluate trees by heuristic search using the tree-bisection reconnection (TBR) branch-swapping algorithm.
The frequencies of synonymous and nonsynonymous substitutions (dS and dN, respectively) were estimated by the Kumar method implemented in the MEGA version 2 program (Kumar et al. 2001) using pairwise alignments of homologous nucleotide sequences from S. glossinidius and Sitophilus zeamais primary endosymbiont (SZPE). Models of selection at the codon level were evaluated using Z-tests (Nei and Kumar 2000), implemented in MEGA2. For Z-tests, the estimates of variance for dN and dS were obtained by sampling 500 bootstrap replicates.
Gene Expression Assays—RNA Isolation
For expression assays, cells from a log-phase culture of S. glossinidius were inoculated into 5 ml growing cultures of Aedes albopictus C6/36 cells at a multiplicity of infection 10. Cultures were maintained at 25°C and harvested for RNA isolation immediately after inoculation of S. glossinidius and at time intervals 4 hours, 8 hours, 24 hours, and 48 hours after inoculation. Beore harvesting, cultures were examined by differential interference contrast (DIC) microscopy to determine the progress of cell invasion and to ensure that the cell cultures were free of contamination. To facilitate the rapid recovery of insect cells and bacteria for RNA isolation, insect cells were detached from culture flasks using a cell scraper, and all cellular material was pelleted immediately at 12,000 x g for 10 min at 25°C. After aspiration of culture media, cell pellets were snap frozen and stored at –70°C before RNA isolation. RNA was isolated from cell pellets using the RNAqueous isolation kit (Ambion, Austin, Tex.), according to the manufacturer's instructions. Isolated RNA was analyzed using the Agilent 2100 nanoanalyzer (Agilent, Palo Alto, Calif.) to confirm the presence of both 16S and 18S rRNA, derived from bacterial and insect total RNA, respectively.
Gene Expression Assays—Quantitative PCR
DNA contamination was removed from the RNA samples by multiple DNase I treatments until no DNA could be detected in TaqMan quantitative PCR assays using primers and probe that detect the S. glossinidius rplB gene. cDNA was prepared from DNA-free RNA samples by reverse transcription in 100 μl reactions (containing 800 ng of RNA) using the random primer approach with TaqMan reverse transcription reagents (Applied Biosystems, Foster City, Calif.).
Primers and probes for the TaqMan flourogenic 5' nuclease assay were designed using Primer Express software (Applied Biosystems) based on the S. glossinidius sequences obtained in the current study (SSR-1, ysaV; SSR-2, and invA) and in a previous study (rplB [Dale et al. 2002]). The sequences and optimal concentrations of the primers and probes used in the TaqMan assays are shown in table 1.
Table 1 Sequences and Optimal Concentrations of TaqMan Primers and Probes Used in This Study
TaqMan assays were conducted in 50 μl reactions using TaqMan Universal PCR Master Mix (Applied Biosystems) in an Applied Biosystems SDS 7000 sequence detection system, under standard cycling conditions (50°C for 2 min, 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 s and annealing/extension at 60°C for 1 min), using a ROX internal standard. The optimum concentration of TaqMan primers and probes (listed in table 1) were determined for each reaction according to the Applied Biosystems SDS 7000 user's manual.
TaqMan assays were conducted in triplicate using each primer/probe combination with 40 ng of cDNA as template. An internal standard curve was generated for each primer/probe combination using serially diluted S. glossinidius DNA as template. Three negative controls were included for each TaqMan assay to assess the integrity of reactions. Relative quantities of transcripts were estimated by the standard curve method (Dale et al. 2002) and the method (Livak and Schmittgen 2001). The use of the method was validated for each primer/probe combination by plotting CT for each primer/probe combination (using rplB as internal control) against log dilution of template DNA. The absolute values of the slopes of all plots were below 0.06, indicating that all primer/probe sets have similar amplification efficiencies (Livak and Schmittgen 2001).
Construction and Characterization of an orgA Mutant
To generate a double-crossover orgA mutant, the complete 1,943-bp prgK-orgA-orgAB sequence was amplified by PCR using primers prgKF (5'-ATGATGATCGCCATGCTGAGC) and orgABR (5'-TAAGGTTCATTTAGCGGC) and cloned into the pGEM T-vector (Promega, Madison, Wis.). The resulting recombinant plasmid DNA was digested with BamHI to cleave the single restriction site at position 46 in the orgA sequence. The linear plasmid was ligated to a kanamycin resistance cassette (Pharmacia, Piscataway, NJ) to generate plasmid pKANorgA. Electrocompetent S. glossinidius were transformed with 20 ng pKANorgA according to methods described previously (Dale et al. 2001). Because the ColE1 origin of replication does not function in S. glossinidius, pKANorgA serves a suicide vector for introduction of the prgK-orgA::kanR-orgAB construct. After electroporation, bacteria were resuspended in 5 ml MM medium and left to recover overnight at 25°C. On the following day, transformed bacteria were subcultured into MM medium containing 20 μg/ml kanamycin to select for recombinants. After overnight incubation in selective media, recombinants were plated onto MM agar containing 20 μg/ml kanamycin and maintained at 25°C in a sealed gas jar under microaerophilic conditions (5% O2, 10% CO2, and 85% N2). Double-crossover mutants were identified by screening kanamycin resistant colonies for ampicillin sensitivity. Kanamycin resistant, ampicillin sensitive recombinants were screened by PCR using primers prgKF and orgABR to confirm the insertional inactivation of orgA.
Invasion Assays
We used time-lapse DIC microscopy to monitor invasion of the C6/36 cell line by wild-type and mutant S. glossinidius. Slide flask monolayer cultures of C6/36 cells were infected with S. glossinidius at low multiplicity of infection (10) and maintained at 25°C. Counts were performed to determine the number of symbionts in host insect cells at time intervals after infection. These counts were performed on at least 25 insect cells within each flask, and the experiment was performed in triplicate to ensure the integrity of data.
Results
BAC Library Screening
Using two sets of PCR primers based on S. glossinidius TTSS sequences obtained during a previous study (Dale et al. 2001), we screened a BAC library constructed from S. glossinidius DNA. The PCR primer sets, known to amplify the invA and spaP homologs from S. glossinidius, each amplified DNA from distinct BAC clones. Complete shotgun sequencing of invA-positive and spaP-positive BAC clones revealed the presence of two TTSS-encoding symbiosis regions (SSR-1 and SSR-2) in different regions of the S. glossinidius genome. The organization of genes within SSR-1 and SSR-2 is shown in figure 1, and the sequences have been deposited in GenBank under accession numbers AY508229 and AY508228.
FIG. 1.— Organization of TTSS genes within the ysa pathogenicity island of Y. enterocolitica, the SSR-1 and SSR-2 regions of S. glossinidius, and the SPI-1 island of S. enterica. Genes are color coded in accordance with their predicted functions based on Blast homology. Putative transcriptional units are highlighted with arrows, indicating both the length and orientation of transcripts. Gene organization is conserved between the Y. enterocolitica ysa island and the S. glossinidius SSR-1 region and differs markedly from the S. glossinidius SSR-2 region and S. enterica SPI-1 island.
S. glossinidius Symbiosis Region 1
The 28.5-kbp SSR-1 is flanked by the rplS and acpM genes on the S. glossinidius chromosome. The current annotation indicates the presence of 23 intact open reading frames and three pseudogenes (defined as regions sharing uncharacteristically low levels of sequence identity with known TTSS genes, having one or more frame-shifting deletions or in-frame stop codons). SSR-1 is defined as a genetic island because it has a base composition (48.9 mol% G+C) substantially lower than the S. glossinidius chromosome (54.9 mol% G+C). In terms of gene organization, SSR-1 most closely resembles the TTSS-encoding ysa pathogenicity island found in Yersinia enterocolitica (Foultier et al. 2002). Both SSR-1 and the ysa island share an almost identical genetic structure and composition, including conservation of major transcriptional units. Because of these similarities, the genes within SSR-1 were named in accordance with their respective ysa homologs. The boundaries of SSR-1 are defined by operons encoding chaperones and effector proteins: the sigD-sigE operon located proximal to rplS and the sycB-yspB-yspC-yspD-yspA operon located proximal to acpM. Based on the available gene inventory, SSR-1 has all the protein components necessary to produce a functional secretion apparatus, including those effector proteins (YspB and YspC) homologous to the Sip proteins of Salmonella enterica that are known to facilitate host cytoskeletal modifications associated with bacterial invasion (Kaniga, Trollinger, and Galan 1995).
S. glossinidius Symbiosis Region 2
The second S. glossinidius symbiosis region, SSR-2, is flanked by the cusC and patA genes on the S. glossinidius chromosome, some 6.5 kbp upstream of the fli operon, encoding flagellar components. The base composition of SSR-2 is 56.3 mol% G+C and is similar to that of the S. glossinidius chromosome (54.9 mol% G+C). Based on ORF identification, GeneMark analysis, and Blast, SSR-2 is predicted to contain 16 intact ORFs and five pseudogenes. Gene organization within SSR-2 resembles that of the Salmonella enterica SPI-1 TTSS island, albeit in a reduced form. Notably, the S. glossinidius SSR-2 has no genes homologous to the sip genes found in S. enterica. In addition, SSR-2 lacks genes homologous to the iacP, iagB, and invH genes that encode an acyl carrier protein, a muramidase, and a chaperone, respectively, in S. enterica. According to the gene inventory, it seems unlikely that secretomes derived from S. glossinidius SSR-2 and S. enterica SPI-1 have comparable function because SSR-2 lacks the sip genes encoding effector proteins that facilitate SPI-1–mediated cell invasion in S. enterica (Kaniga, Trollinger, and Galan 1995). In addition, SSR-2 has no functional homologs of either the S. enterica SPI-1 invE gene, known to positively regulate the secretion of Sip proteins (Kubori and Galan 2002), or the S. enterica SPI-1 invB gene, known to encode a protein chaperone of SipA (Bronstein, Miao, and Miller 2000). Furthermore, SSR-2 lacks functional homologs of those S. enterica SPI-1 genes involved in regulation (spaN [Collazo, Zierler, and Galan 1995; Kubori et al. 2000]) and assembly (prgI, prgJ [Klein, Fahlen, and Jones 2000; Sukhan et al. 2001]) of the TTSS needle substructure. Thus, although SSR-2 maintains all genes necessary to produce the intracellular and membrane-bound components of the TTSS syringe, it lacks functional homologs of genes necessary for the needle substructure.
Ancestry of SSR-1 and SSR-2
Initially, we used parsimony and distance methods to construct TTSS gene trees with homologous sequences derived from a wide range of gram-negative bacteria. Because bootstrap analyses provided little or no support for deep relationships in these trees, we focused on a smaller, well-supported clade from within the initial data set. Subsequent analyses were based on ML approaches, incorporating models of nucleotide substitution derived from hierarchical ML ratio tests (Posada and Crandall 1998). To determine the ancestry of SSR-1 and SSR-2, we constructed gene trees from homologs of the S. enterica invA, invC, and concatenated spaPQR sequences (fig. 2). Only the invA tree was supported by more than 50% of ML bootstrap resamples at every node. The invA homolog from S. glossinidius SSR-1 was placed in a clade supported by 100% of bootstrap resamples with the ysaV gene from Y. enterocolitica and an invA homolog from the SZPE (Dale et al. 2002). The S. glossinidius SSR-2 invA homolog was also placed in a well-supported clade alongside invA sequences from Chromobacterium violaceum and S. enterica. For nodes with more than 50% bootstrap support, the invC tree showed the same overall topology. The S. glossinidius SSR-1 invC homolog was placed in a well-supported clade with the SZPE invC sequence and the Y. enterocolitica ysaN sequence. The S. glossinidius SSR-2 invC homolog was placed in a well-supported clade with the Chromobacterium violaceum and S. enterica invC sequences. Although the spaPQR tree was not as well resolved as the invA and invC trees, the S. glossinidius SSR-1 spaPQR sequence was also placed in a well-supported clade with the Y. enterocolitica ysa sequence, whereas the SSR-2 spaPQR sequence was placed in a clade with the spaPQR sequence from the weevil endosymbiont SZPE. This result is distinct because the SZPE invA and invC sequences were both placed in well-supported clades, along with their respective SSR-1 homologs. During construction of the invA, invC, and spaPQR trees, we included sequences from the E. coli 0157:H7 eiv/epa chromosomal island (O-island #115 [Perna et al. 2001]). This island is notable because, like SSR-2 in S. glossinidius, it also lacks genes encoding the Sip effector proteins. However, the invA, invC, and spaPQR trees do not support a direct ancestry of the E. coli 0157: H7 O-island #115 island and S. glossinidius SSR-2 after the loss of genes encoding the Sip effectors. Rather, it seems more likely that the Sip-encoding genes were lost independently in the lineages leading to S. glossinidius and E. coli 0157. In summary, the results of the phylogenetic analyses lead to three important conclusions. First, sequences from SSR-1 are closely related to homologs found in the Y. enterocolitica ysa island, whereas sequences from SSR-2 are closely related to homologs found in the S. enterica and C. violaceum SPI-1 islands. Second, the inv and spa homologs cloned previously from the weevil endosymbiont SZPE (Dale et al. 2002) most likely originated from two distinct chromosomal regions, analogous to SSR-1 and SSR-2, respectively. Third, the loss of genes encoding Sip effector proteins most likely occurred independently in S. glossinidius and E. coli 0157:H7.
FIG. 2.— Maximum-likelihood trees based on invA, invC, and the concatenated sequences of spaPQR homologs from S. glossinidius, SZPE, and selected enteric pathogens. Trees were generated by an exhaustive ML search process and bootstrapped by heuristic ML search. Bootstrap support values are included for nodes with greater than 50 % support. Trees were rooted with the E. coli flagellar homologs of the TTSS genes.
Genic Components of SSR-1 and SSR-2 Evolve by Purifying Selection
The occurrence of multiple pseudogenes within SSR-2 prompted us to determine the mode of selection operating on certain genes within SSR-1 and SSR-2. To avoid problems associated with saturation in the pairwise estimation of substitution frequencies, we were restricted to focusing on those SSR-1 and SSR-2 gene sequences that have been determined for S. glossinidius and the closely related weevil endosymbiont SZPE (table 2). To determine the modes of selection on these representative SSR-1 and SSR-2 genes, we first used a two-tailed Z-test to evaluate the hypothesis that these genes are evolving neutrally (dN dS). After rejection of this hypothesis for all genes (data not shown), we used a one-tailed Z-test to evaluate the hypothesis that these genes are evolving by purifying selection (dN < dS). This hypothesis was accepted for all genes with strong support (all P values < 0.005), indicating that genes in both SSR-1 and SSR-2 are evolving by purifying selection. Although these data indicate that SSR-2 has retained function despite the loss of genes encoding effectors and protein components of the syringe, we do not exclude the possibility that positive (diversifying) selection is operating locally at some sites within the TTSS-encoding genes.
Table 2 Sequence Divergence Between S. glossinidius and SZPE
Differential Expression of SSR-1 and SSR-2
We used quantitative TaqMan PCR assays to determine the expression profiles of the ysaV gene (located within SSR-1) and the invA gene (located within SSR-2) in S. glossinidius through a time course of infection in Aedes albopictus C6/36 cells. The numbers of transcripts generated from the ysaV and invA genes were measured, along with the numbers of transcripts of a control gene, rplB. Because rplB encodes a ribosomal protein, it is anticipated to be expressed at a constant level in bacteria undertaking translation. In addition, rplB is anticipated to be expressed at a relatively high level because it has a high codon adaptation index in bacteria (Dale et al. 2002; Lithwick and Margalit 2003). The relative abundance of each gene transcript was determined by estimation from an internal standard curve (Dale et al. 2002) and by the method (Livak and Schmittgen 2001). Although both methods of analysis yielded similar results, data generated by the standard-curve method, presented in figure 3, are considered more robust because plots of CT versus log dilution of template DNA produced slightly different slopes for each primer/probe combination tested.
FIG. 3.— Quantitative RT-PCR analysis of gene expression in S. glossinidius at time intervals throughout the initial stages of infection of A. albopictus C6/36 cells. The numbers of transcripts from ysaV, invA, and fusA were measured using the rplB gene as an endogenous control. "Fold-difference" values were determined by comparing transcript numbers in each sample to the 4-hour control sample. Peak levels of the ysaV transcript were detected 24 hours postinfection, when symbionts are actively invading insect cells. Peak levels of the invA transcript were detected 48 hours postinfection, when symbionts were established within host cells.
The expression patterns of ysaV and invA varied with respect to the timing of expression throughout the course of cell invasion. The ysaV gene was maximally expressed 24 hours after infection of the A. albopictus cell line. At this timepoint we observed a greater than 10-fold increase in the expression of ysaV relative to the 4-hour sample. Based on observations conducted in this study and in a previous study (Dale et al. 2001), this timepoint coincides with the invasion of insect cells by S. glossinidius. After invasion, at 48 hours postinfection, we observed a reduction in the numbers of ysaV transcripts, coincident with the establishment of the intracellular stage of infection. In contrast, we observed very little increase in the numbers of invA transcripts at 24 hours after infection of the A. albopictus cell line, when symbionts are most actively engaged in host cell invasion. Instead, invA was maximally expressed at 48 hours postinfection, when bacteria are established within host cells. At this timepoint, we detected a greater than 10-fold increase in the expression of invA relative to the 4-hour sample.
Mutant S. glossinidius Lacking orgA Demonstrate Impaired Replication in Host Cells
To investigate the function of SSR-2 in the process of cell invasion by S. glossinidius, we generated an orgA double-crossover knockout mutant. In S. enterica, the orgA gene is essential for SPI-1–mediated invasion, and the OrgA protein is predicted to be an essential component of the SPI-1 TTSS (Klein, Fahlen, and Jones 2000). Although we identified orgA homologs in both SSR-1 and SSR-2, these genes share a relatively low level of nucleotide sequence homology; hence, the inactivation of the SSR-2 orgA homolog is not expected to affect the function of the SSR-1 TTSS in S. glossinidius. To determine the phenotype of the S. glossinidius orgA mutant, we performed invasion assays in A. albopictus C6/36 monolayer cultures. Both mutant and wild-type S. glossinidius were monitored for their ability to invade and persist in A. albopictus cells over a period of 64 hours after infection (fig. 4). During the first 32 hours after infection, when S. glossinidius is engaged in the invasion of insect cells, we observed no difference in the numbers of orgA mutant and wild-type bacteria invading insect cells. However, after establishment of the intracellular infection, the orgA mutant was substantially impaired in its ability to replicate inside cells, relative to the wild-type strain of S. glossinidius. Taken together with the expression data, these results suggest that SSR-2 has a role distinct from that of SSR-1, enhancing proliferation of S. glossinidius inside insect cells, after SSR-1–mediated entry. In addition, the presence of a clear phenotype in the orgA mutant indicates that mutations leading to the formation of prgI and prgJ pseudogenes, have not eliminated the expression of genes (including orgA) that are located downstream in the putative prgH-orgAb polycistron.
FIG. 4.— Invasion assay in the A. albopictus C6/36 cell line. We counted the numbers of orgA mutant and wild-type S. glossinidius in insect cells at 8-hour intervals throughout the course of infection of C6/36 cells. Although the orgA mutant remains invasive, it has an impaired ability to replicate inside host cells after invasion. Counts were obtained from at least 25 insect cells in three replicate experiments at each timepoint.
Discussion
In the current study, we obtained the complete nucleotide sequences of two chromosomal symbiosis regions, designated SSR-1 and SSR-2, from the mutualistic insect endosymbiont, S. glossinidius. Although SSR-1 and SSR-2 both encode genetic components of type-III secretion, each region has a distinct base composition, gene content, and gene organization, indicative of independent ancestry. Phylogenetic analyses indicate that SSR-1 is most closely related to the ysa pathogenicity island of Y. enterocolitica, whereas SSR-2 is most closely related to the SPI-1 pathogenicity islands found in S. enterica and C. violaceum (Ribeiro de Vasconcelos et al. 2003).
Although TTSS genes have been identified previously in Sodalis and in the closely related weevil endosymbiont SZPE (Dale et al. 2001, 2002), there was no prior indication of the presence of phylogenetically distinct TTSS genes in separate chromosomal regions. However, the phylogenetic analyses presented in the current study indicate that the inv genes obtained in these previous studies were derived from SSR-1, whereas the spa genes were derived from SSR-2. Because it was clearly demonstrated that acquisition of both the inv and spa homologs predates the divergence of a common ancestor of S. glossinidius and SZPE (Dale et al. 2002), we can now assume that SSR-1 and SSR-2 were both present in a presymbiotic ancestor of S. glossinidius and SZPE.
Several bacterial pathogens are known to maintain two or more distinct type-III secretion systems that function in different ways to modulate pathogenesis. For example, S. enterica maintains two TTSS-encoding pathogenicity islands, designated SPI-1 and SPI-2 (Ochman and Groisman 1996). Whereas SPI-1 is expressed upon contact with host cells and facilitates bacterial invasion (Zhou and Galan 2001), SPI-2 is expressed only after invasion and promotes replication of bacteria in host-enclosed vacuoles (Waterman and Holden 2003).
In S. glossinidius, several compelling lines of evidence indicate that SSR-1 has an important role in mediating the invasion of host insect cells. First, SSR-1 maintains all genes necessary to produce a type-III secretome, including those genes encoding secreted effector proteins that facilitate host cytoskeletal modifications associated with invasion. Second, there is a substantial increase in the number of ysaV transcripts in S. glossinidius cells immediately before the invasion of insect cells in vitro. Third, mutant S. glossinidius lacking the SSR-1 ysaN gene are deficient in their ability to invade insect cells in vitro and cannot establish symbiosis in vivo (Dale et al. 2001).
Based on the gene inventory, SSR-2 lacks functional homologs of the spaN, prgI, and prgJ genes necessary for the production of the TTSS needle substructure in S. enterica. Although open reading frames can be detected in the anticipated positions of spaN, prgI, and prgJ, these reading frames are truncated and share uncharacteristically low levels of sequence identity with functional homologs in the public database. This clearly indicates relaxed selection on those genes encoding protein components of the needle. Because our molecular evolutionary analyses indicate that other genes in SSR-2 are evolving under strong purifying selection, it seems that SSR-2 retains functionality despite the absence of intact spaN, prgI, and prgJ genes. Notably, SSR-2 also lacks any genes encoding the Sip effector proteins that are translocated to host cells during SPI-1–mediated invasion by S. enterica. Furthermore, SSR-2 lacks functional copies of the invE and invB genes that are known to facilitate the translocation of Sip proteins in S. enterica (Bronstein, Miao, and Miller 2000; Kubori and Galan 2002). Although the absence of genes encoding effector proteins is striking, the E. coli 0157:H7 pathogen also lacks genes encoding Sip effector proteins within the eiv/epa (O-island #115) (Perna et al. 2001), predicted to encode TTSS proteins. Based on the phylogenetic analyses presented in the current study, the loss of genes encoding the Sip effector proteins in S. glossinidius and E. coli 0157:H7 occurred as a result of independent evolutionary events. Although many pathogens are known to secrete type-III effectors encoded by genes located outside of pathogenicity islands (Cornelis and Van Gijsegem 2000; Waterman and Holden 2003; Chang et al. 2004), these effectors have yet to be identified in S. glossinidius.
Because of the close phylogenetic relationship between genes encoding components of the TTSS in SSR-1 and SSR-2, it is pertinent to consider the possibility of functional complementation between the two secretion systems. Because SSR-2 has a reduced gene inventory relative to SSR-1, one could envisage a scenario in which the missing or inactive genes in SSR-2 are complemented by the respective functional homologs in SSR-1. This observation would imply a process of subfunctionalization (Hughes 1994; Lynch et al. 2001) occurring in SSR-2 as part of a degenerative adaptation to life in the insect host. However, the results of the current study contradict such a hypothesis and suggest instead that SSR-2 has evolved a new and independent function through neofunctionalization (Walsh 2003). This finding is evident from the results of quantitative PCR assays indicating differential expression of transcripts derived from SSR-1 and SSR-2 during invasion. Also, the phenotypes of mutants lacking key components of SSR-1 and SSR-2 are distinct; the SSR-1 ysaN mutant is deficient in its ability to invade insect cells (Dale et al. 2001), and the SSR-2 orgA mutant lacks the ability to replicate inside host cells after invasion.
Although SSR-2 is clearly a descendent of the SPI-1–like pathogenicity islands found in S. enterica and related pathogens, the function of SSR-2 has been modulated in S. glossinidius such that it now more closely resembles the SPI-2 secretion system of S. enterica. Adaptation by gene duplication and neofunctionalization is predicted to represent an important source of innovation in evolution (Ohno 1970). Indeed, these processes have been implicated in the invention of type-III secretion and the origin of pathogenesis in gram-negative bacteria (Horn et al. 2004; Saier 2004). In the context of symbiosis, such innovations could drive important evolutionary transitions, including the transition from parasitism to mutualism (Dale et al. 2001).
Acknowledgements
The authors thank Serap Aksoy (Yale University) for providing a culture of S. glossinidius. We also thank Jon Seger and an anonymous reviewer for providing useful comments.
References
Akman, L., R. V. Rio, C. B. Beard, and S. Aksoy. 2001. Genome size determination and coding capacity of Sodalis glossinidius, an enteric symbiont of tsetse flies, as revealed by hybridization to Escherichia coli gene arrays. J. Bacteriol. 183:4517–4525.
Aksoy, S., X. Chen, and V. Hypsa. 1997. Phylogeny and potential transmission routes of midgut-associated endosymbionts of tsetse (Diptera:Glossinidae). Insect Mol. Biol. 6:183–190.
Bronstein, P. A., E. A. Miao, and S. I. Miller. 2000. InvB is a type III secretion chaperone specific for SspA. J. Bacteriol. 182:6638–6644.
Chang, J. H., A. K. Goel, S. R. Grant, and J. L. Dangl. 2004. Wake of the flood: ascribing functions to the wave of type III effector proteins of phytopathogenic bacteria. Curr. Opin. Microbiol. 7:11–18.
Collazo, C. M., and J. E. Galan. 1997. The invasion-associated type-III protein secretion system in Salmonella: a review. Gene 192:51–59.
Collazo, C. M., M. K. Zierler, and J. E. Galan. 1995. Functional analysis of the Salmonella typhimurium invasion genes invl and invJ and identification of a target of the protein secretion apparatus encoded in the inv locus. Mol. Microbiol. 15:25–38.
Cornelis, G. R., and F. Van Gijsegem. 2000. Assembly and function of type III secretory systems. Annu. Rev. Microbiol. 54:735–774.
Dale, C., and I. Maudlin. 1999. Sodalis gen. nov. and Sodalis glossinidius sp. nov., a microaerophilic secondary endosymbiont of the tsetse fly Glossina morsitans morsitans. Int. J. Syst. Bacteriol. 49:267–275.
Dale, C., G. R. Plague, B. Wang, H. Ochman, and N. A. Moran. 2002. Type III secretion systems and the evolution of mutualistic endosymbiosis. Proc. Natl. Acad. Sci. USA 99:12397–12402.
Dale, C., S. A. Young, D. T. Haydon, and S. C. Welburn. 2001. The insect endosymbiont Sodalis glossinidius utilizes a type III secretion system for cell invasion. Proc. Natl. Acad. Sci. USA. 98:1883–1888.
Foultier, B., P. Troisfontaines, S. Muller, F. R. Opperdoes, and G. R. Cornelis. 2002. Characterization of the ysa pathogenicity locus in the chromosome of Yersinia enterocolitica and phylogeny analysis of type III secretion systems. J. Mol. Evol. 55:37–51.
Gingrich, J. C., D. M. Boehrer, and S. B. Basu. 1996. Partial CviJI digestion as an alternative approach to generate cosmid sublibraries for large-scale sequencing projects. Biotechniques 21:99–104.
Horn, M., A. Ollingro, S. Schmitz-Esser et al. (13 co-authors). 2004. Illuminating the evolutionary history of Chlamydiae. Science 304:728–730.
Hueck, C. J. 1998. Type III protein secretion systems in bacterial pathogens of animals and plants. Microbiol. Mol. Biol. Rev. 62:379–433.
Hughes, A. L. 1994. The evolution of functionally novel proteins after gene duplication. Proc. R. Soc. Lond. B Biol. Sci. 256:119–124.
Kaniga, K., D. Trollinger, and J. E. Galan. 1995. Homologues of the Shigella invasions ipaB and ipaC are required for Salmonella typhimurium entry into cultured cells. J. Bacteriol. 177:3965–3971.
Kelley, J. M., C. E. Fields, M. B. Craven, D. Bocskai, U. Kim, S. D. Rounsley, and M. D. Adams. 1999. High throughput direct end sequencing of BAC clones. Nucleic Acids Res. 27:1539–1546.
Klein, J. R., T. F. Fahlen, and B. D. Jones. 2000. Transcriptional organization and function of invasion genes within Salmonella enterica serovar typhimurium pathogenicity island 1, including the prgH, prgI, prgJ, prgK, orgA, orgB, and orgC genes. Infect. Immun. 68:3368–3376.
Kubori, T., and J. E. Galan. 2002. Salmonella type III secretion-associated protein InvE controls translocation of effector proteins into host cells. J. Bacteriol. 184:4699–4708.
Kubori, T., A. Sukhan, S. I. Aizawa, and J. E. Galan. 2000. Molecular characterization and assembly of the needle complex of the Salmonella typhimurium type III protein secretion system. Proc. Natl. Acad. Sci. USA. 97:10225–10230.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244–1245.
Lithwick, G., and H. Margalit. 2003. Hierarchy of sequence-dependent features associated with prokaryotic translation. Genome Res. 13:2665–2673.
Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-delta delta C(T)) method. Methods 25:402–408.
Lukashin, A., and M. Borodovsky. 1998. GeneMark.hmm: new solutions for gene finding. Nucleic Acids Res. 26:1107–1115.
Lynch, M., M. O'Hely, B. Walsh, and A. Force. 2001. Distribution of pathogenicity islands in Salmonella spp. Genetics 159:1789–1804.
Marie, C., W. J. Broughton, and W. J. Deakin. 2001. Rhizobium type III secretion systems: legume charmers or alarmers? Curr. Opin. Plant Biol. 4:336–342.
Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University Press, New York
Ochman, H., and E. A. Groisman. 1996. Salmonella entry into host cells: the work in concert of type III secreted effector proteins. Infect. Immun. 64:5410–5412.
Ohno, S. 1970. Evolution by gene duplication. Springer, Heidelberg.
Perna, N.T., G. Plunkett, V. Burland et al. (28 co-authors). 2001. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 409:529–533.
Posada, J., and K. A. Crandall. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14:817–818.
Ribeiro de Vasconcelosa, A. T., D. F. de Almeidab, M. Hungriac et al. (112 co-authors). 2003. The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability. Proc. Natl. Acad. Sci. USA 100:11660–11665.
Saier, M. H. 2004. Evolution of bacterial type III protein secretion systems. Trends Microbiol. 12:113–115.
Sukhan, A., T. Kubori, J. Wilson, and J. E. Galan. 2001. Genetic analysis of assembly of the Salmonella enterica serovar typhimurium type III secretion-associated needle complex. J. Bacteriol. 183:1159–1167.
Swofford, D. L. 2000. PAUP*: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, Mass.
Walsh, B. 2003. Population-genetic models of the fates of duplicate genes. Genetica 118:279–294.
Waterman, S. R., and D. W. Holden. 2003. Wake of the flood: ascribing functions to the wave of type III effector proteins of phytopathogenic bacteria. Cell. Microbiol. 5:501–511.
Zhou, D., and J. E. Galan. 2001. Functions and effectors of the Salmonella pathogenicity island 2 type III secretion system. Microbes Infect. 3:1293–1298.(Colin Dale, Tait Jones an)