A tamB homolog is involved in maintenance of cell envelope integrity and stress resistance of Deinococcus radiodurans

The translocation and assembly module (TAM) in bacteria consists of TamA and TamB that form a complex to control the transport and secretion of outer membrane proteins. Herein, we demonstrated that the DR_1462-DR_1461-DR_1460 gene loci on chromosome 1 of Deinococcus radiodurans, which lacks tamA homologs, is a tamB homolog (DR_146T) with two tamB motifs and a DUF490 motif. Mutation of DR_146T resulted in cell envelope peeling and a decrease in resistance to shear stress and osmotic pressure, as well as an increase in oxidative stress resistance, consistent with the phenotype of a surface layer (S-layer) protein SlpA (DR_2577) mutant, demonstrating the involvement of DR_146T in maintenance of cell envelope integrity. The 123 kDa SlpA was absent and only its fragments were present in the cell envelope of DR_146T mutant, suggesting that DR_146T might be involved in maintenance of the S-layer. A mutant lacking the DUF490 motif displayed only a slight alteration in phenotype compared with the wild type, suggesting DUF490 is less important than tamB motif for the function of DR_146T. These findings enhance our understanding of the properties of the multilayered envelope in extremophilic D. radiodurans, as well as the diversity and functions of TAMs in bacteria.


Results
Identification of a tamB-containing locus in D. radiodurans. Two neighbouring sequences (DR_1462 and DR_1461) each containing a tamB motif were identified in the D. radiodurans genome using BLASTP, with each sequence sharing 28% and 29% amino acid sequence identity with Escherichia coli (E. coli) TamB encoded by b4221, respectively. However, a TamA homolog was not detected in D. radiodurans using the E. coli TamA as the query sequence, suggesting D. radiodurans might express a specific TAM lacking TamA.
The current D. radiodurans genome sequence (NCBI accession: NC_001263.1) contains three consecutive annotated ORFs (DR_1462, DR_1461 and DR_1460) in the negative chain of D. radiodurans chromosome 1 (Fig. 1a). By re-sequencing these loci, we found five gaps and four base errors in the DNA sequence between 1463248 and 1475251 ( Supplementary Fig. S1). The gaps introduced invalid start and/or stop codons in the predicted DR_1462, DR_1461 and DR_1460 genes (Supplementary Figs S1 and S2). Therefore, DR_1462, DR_1461 and DR_1460 actually form a single intact ORF, which we refer to as DR_146T (data submitted to NCBI, GenBank accession number is KY352801). This ORF was predicted to encode a TamB homolog of 4002 amino acid residues containing two TamB domains (PFAM signature: PF04357.11) and a C-terminal DUF490 domain, on contrast with E. coli TamB that has one TamB domain and one DUF490 domain ( Fig. 1b and c). A signal peptide of 34 amino acids and a transmembrane helix spanning residues 20-42 were predicted using the SignalIP 3.0 Server and the TMHMM Server v. 2.0, respectively (Fig. 1b). Moreover, homologs of DR_146T are also present as intact hypothetical genes in other sequenced Deinococcus bacteria (Supplementary Table S3). Together these features imply that the DR_146T is a tamB homolog 15 . DR_146T is involved in cell growth in D. radiodurans. The growth of the DR_146T mutant (∆ DR_146T) and the mutant deficient in DUF490 (∆ DR_146T-DUF490) were slower than that of the wild type (Fig. 2). The DR_146T mutant grew twice as slowly as the wild type, while ∆ DR_146T-DUF490 had only a slightly longer doubling time than the wild type. These results indicate that DR_146T might play an important role in cell growth or cell division. Moreover, the growth defect phenotype of ∆ DR_146T is similar to that of the SlpA (DR_2577) mutant (Fig. 2), suggesting that DR_146T might contribute to cell envelope integrity along with the S-layer components. DR_146T is important for cell envelope integrity in D. radiodurans. The wild type colonies were circular and smooth, while the colonies of DR_146T mutant were ring-shaped and rugose similar to those of SlpA mutant ( Supplementary Fig. S3). In liquid media, mutant cells tended to aggregate and settle more easily than wild type cells ( Supplementary Fig. S4). Scanning electron microscopy (SEM) images showed that wild type cells were elliptical and grouped into diplococci or tetracocci (Fig. 3a1), whereas mutant cells displayed surface variation and shedding of the outer layer ( Fig. 3b1 -e1). Meanwhile, the results of transmission electron microscopy (TEM) demonstrated that the ultrastructure of the cell envelope of the mutants was different from that of wild type cells (Fig. 3a2-e2). The cell envelope of ∆ DR_146T exhibited the most severe damage with parts of the outer layer peeling off, and the inner layer exposed to the environment (Fig. 3d). Damage to the cell envelope of ∆ DR_146T-TamB 2nd -DUF490 that is mutated in the second TamB and DUF490 motifs (Fig. 3c) was similar to that of the ∆ DR_146T cells (Fig. 3d), but more severe than that of ∆ DR_146T-DUF490 cells (Fig. 3b), indicating that TamB motif appeared to be more important than the DUF490 motif. SEM and TEM images revealed that the outmost layer peeled in the SlpA mutant (∆ DR_2577) (Fig. 3e). Together with the cell growth phenotype, these results indicated that DR_146T might play an important role in maintaining cell envelope integrity. The cytoplasmic membrane (inner membrane) and peptidoglycan layer could still be formed during cell division, but formation of the outer layers was clearly inhibited in the mutant.
Mutation of DR_146T alters stress resistance in D. radiodurans. Under continuous vortexing, the survival fraction of ∆ DR_146T was substantially lower than that of the wild type ( Fig. 4), consistent with the survival phenotype of the SlpA mutant 28 . The survival of ∆ DR_146T-DUF490 was slightly lower than that of the wild type. Fig. 5 shows that the cell resistance of ∆ DR_146T and ∆ DR_146T-DUF490 to osmotic pressure was lower than that of the wild type, indicating that mutation-induced deficiency in cell envelope integrity led to higher sensitivity to osmotic pressure. These suggest that DR_146T contributes to stress resistance, and the DUF490 motif might be less important than the TamB motif for the function of DR_146T.
The resistance of ∆ DR_146T and the SlpA mutant (∆ DR_2577) to oxidative stress was much higher than that of wild type cells ( Fig. 6a and b). The ∆ DR_146T strain could survive high concentrations of H 2 O 2 (160 mM) without any obvious decline in viability compared with controls (0 mM H 2 O 2 ), indicating that disruption of DR_146T leads to an increase in resistance to oxidative stress. Previous studies demonstrated that accumulation of manganese ions (Mn 2+ ) and a higher intracellular Mn/Fe ratio contribute to oxidative stress resistance in D. radiodurans through Mn complex-mediated ROS scavenging 25,26,34 . Thus, we measured the metal ion content in mutant and wild type cells by inductively-coupled plasma-mass spectrometry (ICP-MS) ( Fig. 6c and d). The Mn ion content in ∆ DR_146T and ∆ DR_2577 was almost twice that of wild type cells, while the Fe and Zn ion content were slightly reduced in the mutants. Therefore, the H 2 O 2 resistance of ∆ DR_146T might be attributed to the increased Mn/Fe ratio in the mutant cells.
Altered proteins in cell envelope, peeling fraction and culture's supernatant. Each of the obtained whole cell envelope, culture's supernatant and peeling cell envelope fractions was divided in half: half was used for mass spectrometry (MS) analysis, and the other half was for SDS-PAGE analysis on protein patterns of the wild type and the mutants. More than 80 proteins including DR_146T (matched to the predicted DR_1462, DR_1461 and DR_1460 from the NCBI database) were detected in the cell envelope of D. radiodurans wild type using MS analysis (Supplementary Table S4). Furthermore, the detection of peptides matched to the DR_146T from the cell envelope of D. radiodurans wild type cells by MS ( Supplementary Fig. S5) suggests that DR_146T is a component of the cell envelope.  The cell envelope of ∆ DR_146T exhibited severe damage with parts of the outer layer peeling off, and the inner layer exposed to the environment compared with that of the wild type. Images represent the SEM and TEM results for D. radiodurans wild type (a1-a2), ∆ DR_146T-DUF490 (b1-b2), ∆ DR_146T-TamB 2nd -DUF490 (c1-c2), ∆ DR_146T (d1-d2) and ∆ DR_2577 (e1-e2), respectively. The inset diagram in (b2) shows an amplified region of the cell envelope. Scale bars indicate the corresponding length.  Table S5). Compared with wild type cells, many proteins including the 123 kDa SlpA were absent from the cell envelope fraction of ∆ DR_146T ( Fig. 7b and c), suggesting that DR_146T might be involved in maintenance of the surface layer. However, the cell envelope protein profile in ∆ DR_146T-DUF490 was similar to that of the wild type, confirming that DUF490 may be not a pivotal factor in maintenance of the cell surface layer.
To further probe the roles of DR_146T in cell envelope assembly and integrity, proteins in the peeling cell envelope and culture's supernatant were identified by MS, respectively. More than 30 proteins were identified in the culture's supernatant of ∆ DR_146T (Table 1), while only a few extracellular nucleases and proteases were detected in the culture's supernatant of wild type cells (data not shown). Among these, surface layer proteins including Hpi (DR_2508), putative S-layer-like array-related protein (DR_1115, DR_1185, DR_0383 and DR_1124) and putative periplasm-located proteins (DR_1571, DR_1027, DR_0363 and DR_A0210) were detected in the culture's supernatant of ∆ DR_146T, suggesting some cell envelope proteins were released into the supernatant following mutation of DR_146T. More than 50 proteins were identified in the peeling envelope of ∆ DR_146T (Table 2), and a number of proteins detected in the culture's supernatant were also found in the peeling envelope fraction. However, some proteins were exclusive to the peeling envelope fraction, including SlpA, the putative outer membrane protein BamA (DR_0379) and the secretin (DR_0774) 31 . SlpA was one of the most abundant proteins in the peeling fraction but was absent in the culture's supernatant, possibly due to the presence of its SLH domain which keeps it anchored to the outer membrane 32,35 . Since the 123 kDa SlpA was not detected in the cell envelope of ∆ DR_146T (Fig. 7c) and proteins identified by MS were based on peptide matches following enzyme digestion, the SlpA detected in the peeling fraction could be fragments of SlpA. These results indicated that many periplasmic, outer membrane and S-layer proteins or their fragments dissociated from cells and were released into the culture in the absence of DR_146T. Proper assembly of the S-layer might be inhibited by mutation of this special TamB homolog in D. radiodurans.
Relationship between DR_146T and S-layer protein SlpA. SDS-PAGE revealed that a protein band of 123 kDa corresponding to SlpA was missing from the envelope of ∆ DR_146T (Fig. 7c). The expression level and location of SlpA were investigated using western blotting. The 123 kDa SlpA was detected in whole-cell extracts and cell envelope fractions of the wild type and the ∆ DR_146T-DUF490 cells ( Fig. 7d and f). By contrast, SlpA was present only as 20 kDa and/or 40 kDa fragments in the peeling fraction and whole cell envelope of ∆ DR_146T ( Fig. 7e and f).

Discussion
We identified the gene DR_146T as a tamB homolog in the extremophilic bacterium D. radiodurans, which is known for its multilayered cell envelope and exceptional stress resistance. Detection of homologs of DR_146T in other Deinococcus species indicates that the gene is conserved in this genus.
Evidence from several lines suggests that DR_146T is an intact functional gene. Firstly, we re-sequenced the gene loci containing the putative tamB homologs and found five gaps and four base errors that led to a frame shift and introduced invalid start and/or stop codons in the predicted DR_1462, DR_1461 and DR_1460 sequences. Analysis of the corrected sequences confirmed the presence of an intact ORF. Secondly, homologs of DR_146T are also present as intact ORFs in other sequenced Deinococcus bacteria. Thirdly, and conclusively, peptides of DR_146T were detected in the cell envelope of D. radiodurans wild type cells by MS analysis (Supplementary Fig. S5). A schematic diagram of the multilayered cell envelope of D. radiodurans is proposed in Fig. 8, adapted from previous reports 28,31,36,37 . DR_146T might function as a TamB-like protein, anchored in the inner membrane with its N-terminal transmembrane helix and spanning the peptidoglycan and periplasm, based on the topology of TamB in E. coli 11 . However, the unusual DR_146T containing two TamB domains is larger than typical bacterial TamB which contains only a single TamB domain 15 , such as TamB in E. coli, Citrobacter rodentium and Salmonella enterica 10 . TamB in the inner membrane is reported to cooperate with TamA in the outer membrane to form a hetero-oligomeric TAM complex 10,11 . However, there does not appear to be a TamA homolog in D. radiodurans, indicating that DR_146T might act as a solo TamB. The presence of TamB in other bacteria lacking TamA has been reported, and the TAM system is suggested to have evolved from an original combination of TamB and BamA, which evolved into TamA in Proteobacteria by a later gene duplication event 15,29,38 . Whether there was any interaction between TamB and BamA in D. radiodurans is not clear. We did not rule out the possibility that other genes with no or little homology to tamA might have similar functions to tamA in D. radiodurans. Considering that the thickness of the cell envelope of D. radiodurans is at least twice that of E. coli 39 , the two TamB domains might be needed for DR_146T to span the thicker peptidoglycan and periplasm in D. radiodurans.
In DR_146T mutant cells, the cell envelope was disrupted and some periplasmic and S-layer proteins were released, suggesting DR_146T is involved in maintaining the integrity of the cell envelope. A number of periplasmic and S-layer proteins were detected at altered relative abundance in the peeling cell envelope of DR_146T mutants, indicating that the proteome integrity of the cell envelope was altered by the mutation. Recently, Smith et al. reported that the abundance of 12 proteins, including protein quality control systems and protein secretion, were altered in the membrane of the TamB homolog (MorC) mutant strain of Aggregatibacter actinomycetemcomitans compared with the wild type strain 40 . The main functions of the S-layer are maintaining cell shape, acting as a mechanical barrier 3,4,7 , providing a binding scaffold for large molecules and ions 41 , and mediating bacterial adhesion 42,43 . DR_146T might facilitate the assembly of the S-layer, because its mutation results in the loss of the 123 kDa SlpA, which is involved in formation and attachment of the surface layer to the inner cell envelope 28 . In addition, the S-layer functions as a sheath surrounding groups of cells, and forming on the surface of daughter cells when they separate 44 . In the absence of DR_146T, detachment of the S-layer sheath from the cell surface was observed in SEM and TEM images (Fig. 3). The S-layer protein of Thermus thermophilus, an ancient bacterial lineage closely related to Deinococcus, was assembled under the assistance of the BAM complex 33 . The S-layer of D. radiodurans may be assembled under the assistance of TamB encoded by DR_146T.
We demonstrated that deficiencies in the cell envelope following mutation of DR_146T led to increased sensitivity to shear stress and osmotic pressure due to a decrease in rigidity of the cell envelope, consistent with the SlpA mutant. Meanwhile, DR_146T and SlpA mutants showed increased resistance to oxidative stress, as well as an increased intracellular Mn 2+ content and Mn/Fe ion ratio, both of which contribute to oxidative stress resistance through non-enzymic ROS scavenging 25,26,45 . An increase in the permeability of the cell envelope in the mutant strains might facilitate the penetration of Mn ions across the cell wall. Furthermore, we cannot eliminate the possibility that mutation of DR_146T affected metal ion transporters directly. Indeed, an iron ABC transporter substrate-binding protein (DR_B0125) was detected in the peeling cell envelope of the DR_146T mutant, suggesting that it was detached, and this could explain the observed decrease in Fe ion content in the mutants. The mechanism by which DR_146T is involved in intracellular Mn accumulation remains to be explained by further studies. In conclusion, tamB homolog DR_146T in D. radiodurans consists of two tamB motifs and one DUF490 motif. DR_146T might be involved in the maintenance of cell envelope integrity and stress resistance through its impact on SlpA. Further investigations are necessary to elucidate protein conformation, structure-activity relationships and details of the mechanism of this unusual TamB homolog in cell envelope assembly. Given the important roles of the TamB homolog in cell envelope integrity, these findings are not only useful for understanding the mechanism of TAM in this and other organisms, but also valuable for screening new antibiotics that target TamB.

Methods
Bacterial strains and growth conditions. All bacterial strains, plasmids, and primers used in this study are listed in Supplementary Tables S1 and S2, respectively. D. radiodurans was cultured in TGY liquid medium (0.5% tryptone, 0.3% yeast extract, 0.1% D-glucose) or on TGY plates (TGY liquid medium supplemented with 1% agar) at 30 °C. E. coli was grown in LB broth (1% tryptone, 0.5% yeast extract, 1% NaCl) or on LB plates (LB liquid medium supplemented with 1% agar) at 37 °C. Appropriate antibiotics were added into the medium where required.
Sequencing analysis of gene loci containing tamB motifs. Gene loci on chromosome 1 (NC_001263.1) containing tamB sequences were searched using BLAST (NCBI, http://blast.ncbi.nlm.nih.gov/ Blast.cgi) with tamB from E. coli as the query sequence. Target DNA was sequenced using a 3730xl DNA Analyzer (ABI, USA). ORFs were predicted using OFR finder (https://www.ncbi.nlm.nih.gov/orffinder/). DNA and protein sequence analyses were performed using BLAST. Signal peptide analysis was conducted using the SignalIP 3.0 Server (http://www.cbs.dtu.dk/services/SignalP-3.0/). Transmembrane helices were analysed using the TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM-2.0/). Construction of mutant strains. Tripartite ligation and double-crossover recombination methods were used as described previously 46,47 with some modifications. Briefly, DNA fragments upstream and downstream of the targeted sequence were amplified and digested with BamHI and HindIII, respectively. The streptomycin resistance fragment was amplified from vector pMD18-T and digested with BamHI and HindIII. Upstream, downstream and streptomycin resistance fragments were ligated with T4 DNA ligase, and ligation products were transformed into competent D. radiodurans cells using the CaCl 2 method. The homozygous mutant strain was confirmed by PCR and DNA sequencing.
Growth curves of mutant and wild type strains. Wild type and mutant strains were cultivated to an OD 600 of ~1.0, inoculated into fresh TGY medium at a ratio of 1:50 (v/v) and incubated at 30 °C. Bacterial growth was monitored by measuring the OD 600 value using a SpectraMax M5 microplate reader (Molecular Devices, USA). Experiments were performed independently in triplicate.

Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) analyses.
All experimental bacterial strains were cultured to an OD 600 of ~1.0. For SEM analysis, harvested cells were double fixed and dehydrated as described by Li 48 , and dehydrated cells were coated with gold-palladium and observed using a SU8010 ultra-high resolution SEM (Hitachi, Japan).
For TEM analysis, specimens were processed as described by Li 48 , cut into ultrathin (70-90 nm) sections using microtome, stained with uranyl acetate and alkaline lead citrate for 15 min, and observed using a Hitachi H-7650 TEM (Hitachi, Japan). Survival assays under shear stress, osmotic pressure and oxidative stress. Bacterial survival assays under shear stress were performed as described previously 28 with some modifications. Briefly, bacterial cultures at an OD 600 of ~1.0 were diluted with fresh TGY medium to OD 600 = 0.1, mixed with sterile zirconia ceramic beads, and exposed to shear stress using a tissue homogenizer (RETSCH MM301, Germany) at 30 Hz for different treatment time (0, 2, 4 or 8 min). Suspensions were diluted and spread onto TGY plates. Bacterial colonies were counted and survival fraction was measured. For the osmotic pressure experiment, bacterial cultures at an OD 600 of ~1.0 were inoculated into fresh TGY medium to which NaCl (0.05 M, 0.1 M, 0.2 M and 0.5 M) was added. Cultures were grown with shaking at 30 °C and the OD 600 was measured using a SpectraMax M5 microplate reader. For the oxidative stress experiment, bacteria were treated with H 2 O 2 in sterile phosphate buffer (0.01 M, pH 7.4) for 30 min, diluted and spread onto TGY plates. For the dripping test, 6 μ l of cells was dripped onto TGY plates. All experiments were independently performed in triplicate.

Measurement of intracellular metal ion concentration by inductively-coupled plasma-mass spectrometry (ICP-MS).
ICP-MS assays were performed as described previously by Sun 34 with some modifications. D. radiodurans wild type and mutant strains were cultured to an OD 600 of ~1.0. Cells were collected and    The TAM complex containing the DR_146T and an unknown cooperating protein was indicated by dashed box. In wild type cells, the S-layer assembles properly and acts as a scaffold for the cell envelope. In the DR_146T mutant, SlpA (DR_2577) is present as fragments in the peeling cell envelope. Components in the carbohydrate coat, outer membrane and periplasm dissociate from the cell.
Peeling fractions were washed three times with 0.01 M phosphate buffer (pH 7.4). During sample preparation, protease inhibitor cocktail (Selleckchem, USA) was added to protect proteins against proteolytic degradation.
Each of the obtained whole cell envelope, culture's supernatant and peeling cell envelope fractions was divided in half: half was used for MS analysis, and the other half was for SDS-PAGE analyses on protein patterns of the wild type and the mutants. For electrophoresis, 6-10% (w/v) separating gels and 5% (w/v) stacking gels were used. The gels were stained with Coomassie Brilliant Blue G250.

Identification of proteins by MS.
Protein samples in solution were denatured by RapiGest SF (Waters, USA), reduced with TRIS-(2-carboxyethyl)-phosphine and alkylated with 20 mM iodoacetamide for 45 min at room temperature in darkness. Trypsin was added at 1:50 trypsin-to-protein mass ratio for digestion overnight and 1:100 trypsin-to-protein mass ratio for a second 4 h-digestion. The peptides were then desalted and dried by vacuum evaporation. For proteins in selected gel slice, the gel was diced into small pieces and reduced with TRIS-(2-carboxyethyl)-phosphine, then alkylated with 20 mM iodoacetamide for 45 min at room temperature in darkness. The obtained samples were placed into 0.65 mL siliconized tubes, then washed with 100 μ l 50% acetonitrile/25 mM NH 4 HCO 3 for 3 times and dried by vacuum evaporation. After overnight trypsin digestion, 5 μ l of 5% formic acid was added to the sample to stop the reaction. The digested solution was transferred into a clean 0.65 ml siliconized tube. 50 μ l 50% acetonitrile/5% formic acid was added to wash the peptides for 3 times, and the peptide supernatant was dried by vacuum evaporation. Then the peptides were diluted with 0.1% formic acid.
The peptides were analysed by HPLC-MS/MS (Triple TOF 5600 + LC/MS/MS system, AB SCIEX, USA) as described previously 28,31 with some modifications. Samples were loaded on a prepacked column (200 μ m × 500 mm, Chrome XP C18-CL, 3 μ m, 300 Å) to desalinate for 10 min at a flow rate of 4 μ l/min. Subsequently, they were eluted at 0.3 μ l/min with a C18 HPLC reversed-phase column (75 μ m × 150 mm, Chrome XP C18-CL, 3 μ m, 300 Å) using a mobile phase of solution A (5% acetonitrile and 0.1% formic acid) and solution B (95% acetonitrile and 0.1% formic acid). During elution, solution B was increased linearly from 5% to 40% over 60 min. Eluted peptides were introduced directly into the mass spectrometer and peptide identification was performed using the D. radiodurans protein sequence database (NCBI). The unweighted spectrum count of each protein was divided by its mass, and the resulting index was used as an indicator of relative protein abundance in each sample 31 . Western blotting. Protein expression levels were confirmed using western blotting as described previously 49 .
A 6× His tag was fused to the C-terminal of SlpA (DR_2577), and monoclonal anti-6× His mouse antibody (Protein tech, USA) was used to detect SlpA-6× His. Horseradish peroxidase-conjugated goat anti-mouse and anti-rabbit IgG were added as secondary antibodies. The expression level of GroEL served as an internal loading control, and was detected using rabbit anti-GroEL polyclonal antibody (Sigma, USA). Statistical analysis. Data were processed using SPSS 18.0 statistical software (SPSS, USA) and are presented as means ± SD. F-tests and independent sample T-tests were used to assess the significance of differences between results. P < 0.05 was considered statistically significant.