|Home | About | Journals | Submit | Contact Us | Français|
Bacterial cell division is an essential process driven by the formation of a Z-ring structure, as a cytoskeletal scaffold at the mid-cell, followed by the recruitment of various proteins which form the divisome. The cell division interactome reflects the complement of different interactions between all divisome proteins. To date, only two cell division interactomes have been characterized, in Escherichia coli and in Streptococcus pneumoniae. The cell divison proteins encoded by Neisseria gonorrhoeae include FtsZ, FtsA, ZipA, FtsK, FtsQ, FtsI, FtsW, and FtsN. The purpose of the present study was to characterize the cell division interactome of N. gonorrhoeae using several different methods to identify protein-protein interactions. We also characterized the specific subdomains of FtsA implicated in interactions with FtsZ, FtsQ, FtsN and FtsW.
Using a combination of bacterial two-hybrid (B2H), glutathione S-transferase (GST) pull-down assays, and surface plasmon resonance (SPR), nine interactions were observed among the eight gonococcal cell division proteins tested. ZipA did not interact with any other cell division proteins. Comparisons of the N. gonorrhoeae cell division interactome with the published interactomes from E. coli and S. pneumoniae indicated that FtsA-FtsZ and FtsZ-FtsK interactions were common to all three species. FtsA-FtsW and FtsK-FtsN interactions were only present in N. gonorrhoeae. The 2A and 2B subdomains of FtsANg were involved in interactions with FtsQ, FtsZ, and FtsN, and the 2A subdomain was involved in interaction with FtsW.
Results from this research indicate that N. gonorrhoeae has a distinctive cell division interactome as compared with other microorganisms.
The online version of this article (10.1186/s12866-017-1140-1) contains supplementary material, which is available to authorized users.
Cell division is essential for bacterial survival. In Escherichia coli (Ec), normal cell division is driven by the formation of an FtsZ-ring at the division site , followed by the recruitment of other essential proteins, which together form the divisome . Genes encoding most cell division proteins are located in a conserved region, the division and cell wall (dcw) cluster . dcw clusters have been identified in most bacterial species, including E. coli, Bacillus subtilis (Bs), Streptococcus pneumoniae (Sp), Caulobacter crescentus (Cc) and Neisseria gonorrhoeae (Ng) [4–7]. Although the gene organization of the dcw cluster varies in different bacteria species , proteins involved in the cell division process are relatively conserved [9, 10].
E. coli encodes ten essential cell division proteins, including FtsZ, FtsA, ZipA, FtsK, FtsQ, FtsB, FtsL, FtsW, FtsI, and FtsN [11, 12]. Assembly of the FtsZ-ring structure is initiated with the polymerization of FtsZ, driven by GTP hydrolysis, at the mid-cell . FtsA and ZipA are recruited by FtsZ and anchor FtsZ to the inner membrane . After the recruitment of FtsK, a DNA translocase involved in DNA segregation [15–17], the protein complexes FtsQ-FtsB-FtsL and FtsW-FtsI are localized to the septal ring, sequentially [15, 18]. Recent studies showed that the FtsQ-FtsB-FtsL complex serves as a signal sensor which promotes cell wall remodeling necessary for cell constriction . FtsI is a high-molecular-weight transpeptidase that cross-links glycan strands. The FtsW-FtsI complex is part of the peptidoglycan synthesis machinery, and FtsW, a lipid II flippase, transports the cell wall precursor across the membrane [20, 21]. FtsN is recruited as the last essential division protein that initiates cell constriction .
Using a bacterial two-hybrid (B2H) assay, an E. coli cell division protein-protein interaction network, the cell division interactome, which included 16 interactions between 10 cell divison proteins, was identified [23, 24]. The cell division interactome of S. pneumoniae was also characterized using a combination of B2H and co-immunoprecipitation assays . A total of 17 interactions was observed among nine cell division proteins of S. pneumoniae which included FtsZ, FtsA, FtsK, DivlB, DivlC, FtsL, FtsW, and PBP2x . To date, E. coli and S. pneumoniae are the only two organisms with characterized cell division interactomes [23–25].
N. gonorrhoeae is a Gram-negative diplococcus that causes gonorrhea in humans . Previous studies on N. gonorrhoeae cell division focused on its Min system which localizes FtsZ to the mid-cell, and FtsZ [27–29]. N. gonorrhoeae also contains a dcw cluster which encodes 5 cell division proteins - FtsZ, FtsA, FtsQ, FtsW, and FtsI . Other non-dcw cluster divisome proteins encoded by N. gonorrhoeae include ZipA, FtsK, and FtsN. As compared to E. coli, N. gonorrhoeae lacks FtsB and FtsL .
To investigate the cell division interactome in N. gonorrhoeae, its cell division protein interactions were identified using a combination of B2H and glutathione S-transferase (GST) pull-down assays, as well as surface plasmon resonance (SPR). We identified nine interactions among the eight cell division proteins tested. We also identified the subdomains of FtsANg involved in its interaction with FtsQNg, FtsZNg, FtsNNg, and FtsWNg. Comparison of the cell division interactomes of E. coli, S. pneumoniae and N. gonorrhoeae indicates that N. gonorrhoeae possesses a distinctive cell division interactome.
The bacterial strains and plasmids used in this study are shown in Table 1. E. coli DH5α and XL1-Blue were used as hosts for cloning. E. coli BL21 (DE3) and C41 (DE3) were used as hosts for protein purification. E. coli R721 was used in B2H assays . E. coli DH5α, XL1-Blue, BL21(DE3) and C41 (DE3) were grown in Luria-Bertani (LB) medium (BD Difco™, Sparks, MD), for 16–18 h (hr), at 37 °C. E. coli R721 was grown under the same conditions and incubated at 34 °C, as described previously .
N. gonorrhoeae CH811 was grown on GC medium base agar (GCMB, Oakville, ON), supplemented with Kellogg’s defined supplement (GCMBK, 40 g D-glucose, 1 g glutamine, 10 ml of 0.5% ferric nitrate and 1 ml of 20% cocarboxylase), at 35 °C, in a humid environment, with 5% CO2, for 18 to 24 h .
When required, the following concentrations of antibiotics were added to LB medium: 100 μg/ml ampicillin (Sigma, Oakville, ON) or 50 μg/ml kanamycin (Sigma). For B2H assays, 34 μg/ml chloramphenicol (Sigma), 30 μg/ml kanamycin, and 50 μg/ml ampicillin were added to LB medium.
N. gonorrhoeae CH811 genomic DNA was purified using a QIAamp® genomic DNA kit (Qiagen, Mississauga, Ontario, Canada). DNA samples were stored at −20 °C. Oligonucleotides for polymerase chain reaction (PCR) amplifications were synthesized by Invitrogen (Table 2; Burlington, Ontario, Canada). PCRs were performed in a GeneAmp® PCR system 9700 (Applied Biosystems, Foster City, CA, USA) as follows: 4 min (min) at 94 °C, 30 cycles of denaturation for 1 min at 94 °C, annealing for 45 s (s) at 55 °C, extension for 1.5 mins at 72 °C, and 10 mins at 72 °C. PCRs were carried out in 100-μl (final volume) mixtures comprising 71.5 μl double-distilled H2O (ddH2O), 10 μl of 10× PCR buffer [15 mM MgCl2, 4 μl of 10 mM deoxynucleoside triphosphate (dNTP), 2 μl of each primer (0.2 μg/ml), 0.5 μl of Taq DNA polymerase (5 U/μl; New England BioLabs, Ontario, Canada)], and, 10 μl of purified N. gonorrhoeae CH811 genomic DNA suspension.
The method developed by Di Lallo et al.  was used for all B2H assays. ftsA, ftsK, ftsQ, ftsI, ftsW, and ftsN were amplified from N. gonorrhoeae CH811 by PCR using the primer pairs P1/P2, P3/P4, P5/P6, P7/P8, P9/P10, and P11/P12 (Table (Table2),2), respectively. PCR amplicons were digested with BamHI and SalI and ligated into previously digested pcIp22 and pcI434 vectors, to produce pclp22-A, pcIp22-K, pcIp22-I, pcIp22-W, pcIp22-Q, pcIp22-N, pcI434-A, pcI434-K, pcI434-I, pcI434-W, pcI434-Q, and pcI434-N (Table 3). zipA Ng was amplified from N. gonorrhoeae CH811 genomic DNA using the primer pair P13/P14 (Table (Table2);2); the PCR amplicons was digested with BglII and BamHI, and ligated into pre-digested pcIp22 and pcI434 to produce pcIp22-ZipA and pcI434-ZipA. pcIp22-Z and pcI434-Z constructs were generated previously .
The expression of ftsA Ng, ftsZ Ng and zipA Ng from B2H constructs was verified by Western blot analysis using appropriate antibodies prepared in our lab using previously described methods . These proteins were expressed from the vectors under the conditions tested (data not shown). The expression of these proteins indicated that any negative B2H interactions involving them was not a function of lack of expression.
To ascertain what subdomains of FtsANg interacted with gonococcal cell division proteins FtsZNg, FtsQNg, FtsWNg, or FtsNNg, six previously created truncations of the protein (T1, T2, T3, T4, T5, and T6; Additional file 1: Figure S1) were used . Plasmid constructs for B2H assays were previously generated .
B2H assays were performed as described previously . This assay is based on the reconstitution of a chimeric repressor that binds to the 434/P22 hybrid operator and represses the expression of a downstream lacZ gene in E. coli R721. Each gene tested for a potential interaction was cloned into pcIp22 and pcI434 and recombinant constructs were transformed into E. coli R721 either singly or in combination. N. gonorrhoeae FtsZ self-interaction was used as positive control. R721 without plasmids and single plasmid transformants were used as negative controls. R721 without plasmids had a β-galactosidase activity of 2504±34 Miller units. The β-galactosidase activity of each combination was compared to that of R721. Values of less than 50% (<1250 Miller Units) indicate a positive interaction between two proteins, while values of more than 50% (>1250 Miller Units) indicate a negative interaction . Statistical analyses were performed using the unpaired Student t-test. Standard deviations were determined for the mean value of Miller units where three independent experiments were performed.
For His-fusion constructs, full-length ftsA, ftsQ, and ftsZ were PCR-amplified from N. gonorrhoeae CH811 genomic DNA using primer pairs P15/P16, P17/P18 and P19/P20 (Table (Table2),2), respectively. PCR amplicons were digested with EcoRI and BglII and ligated into pre-digested pET30a, to create pETA, pETQ, and pETZ. Plasmid pETA was transformed into E. coli C41 (DE3) and plasmids pETQ and pETZ were transformed into E. coli BL21 (DE3). The overexpression of all fusion proteins was induced with 400 μM IPTG, at 30 °C, for 2 h. Purification of His-FtsANg, His-FtsZNg, and His-FtsQNg was completed using His•Bind® Resin (EMD Millipore, Billerica, MA), following the manufacturer’s instructions. His-FtsZNg was further treated with thrombin protease (EMD Millipore, Billerica, MA), overnight, at 4 °C, to cleave the N-terminal His tag. Thrombin was removed using 100 μl of p-aminobenzamidine-agarose (Sigma #A7155). FtsZ was dialyzed against MES buffer (50 mM MES, 300 mM KCl, 10 mM MgCl2, pH 7.5) prior to use in FtsZ polymerization experiments .
For GST fusion constructs, full-length ftsA and ftsN were PCR-amplified, from N. gonorrhoeae CH811, using primer pairs P21/P18 and P22/P23 (Table (Table2),2), respectively. The ftsA amplicon was digested with BamHI and EcoRI and ligated into pre-digested pGEX2T, to create pGEXA (Table (Table3).3). The ftsN amplicon was digested with EcoRI and XhoI and ligated into pre-digested pGEX2T, producing pGEXN (Table (Table3).3). Plasmids pGEXA and pGEXN were transformed into E. coli C41 (DE3) and E. coli BL21 (DE), respectively. Overexpression of GST-FtsA and GST-FtsN was accomplished by induction with either 400 μM or 800 μM of IPTG, respectively, at 30 °C, for 2 h. Purification of GST-FtsA and GST-FtsN was carried out using GST•Bind™ Resin (EMD Millipore, Billerica, MA), following the manufacturer’s instructions.
Purified GST-fusion and His-fusion proteins were incubated with pre-equilibrated GST•Bind™ Resin in phosphate buffered saline (PBS) buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, 0.5% Triton-X100, 1 mM DTT, pH 7.9) at 4 °C overnight. Pre-purified GST was used as a negative control. The pre-bound resin was collected by centrifugation and washed in PBS three times. Bound proteins were dissociated from resin by adding 5X Laemmli buffer, separated by electrophoresis on 10% sodium dodecyl sulfate polyacrylamide gels (SDS-PAGE), and identified by Western blot using polyclonal anti-GST or anti-6×His antibodies (Thermo Scientific; Waltham, MA), sequentially.
For FtsANg-FtsZNg interactions, the GST pull-down assay was performed in MES buffer (50 mM MES-NaOH, 50 mM KCl, 10 mM MgCl2, 0.5% Triton-X100, 1 mM ATP, 2 mM GTP, pH 7.5) . To promote the polymerization of FtsZNg necessary for this interaction, FtsZNg was treated with 2 mM GTP and 1 mM ATP, as described previously , before mixing with GST-FtsANg and GST•Bind™ Resin.
All GST pull-down assays were performed minimally in duplicate.
FtsZNg polymerization was measured by 90° angle light scattering using a Dynapro-MS800 instrument (Wyatt Technology Corporation) with a wavelength of 310 nm and a slit width of 0.5 mm. MES buffer is optimal for FtsZ polymerization which is required to observe an FtsA-FtsZ interaction [34, 35]. FtsZNg (~6 μM) in MES buffer (50 mM MES-NaOH, 50 mM KCl, 10 mM MgCl2, pH 7.5) was injected into a 45 ul quartz cuvette and warmed to 30 °C, prior to the measurement. Data were collected, for 4 min, from unpolymerized FtsZNg to establish a baseline. GTP was then added to a final concentration of 2 mM and data were collected every 5 s for 25 min. Data were recorded and analyzed using Dynamics v5 software.
Negative stain electron microscopy was used to visualize FtsZNg polymers. 5 μl of FtsZ (6 μM) with, or without, GTP (final concentration 2 mM) was incubated, at 30 °C, for 5 min. The mixture was placed on a carbon-coated copper grid (400 mesh size) for 2 min and then blot dried. The grid containing FtsZNg was stained with 1% uranyl acetate, blotted, and air-dried for 3 h. Polymers were visualized and photographed using a Hitachi transmission electron microscopy HT7700.
Protein interactions were examined by SPR using a Bio-Rad XPR36 (Bio-Rad Laboratories) instrument and a ProteOn™ HTE Sensor Chip (Bio-Rad Laboratories). The chip surface was regenerated by injection of 0.5% SDS, 50 mM NaOH, 100 mM HCl and 300 mM EDTA, at a flow rate of 30 μl/min, for 120 s. Activation was performed using 500 μM of NiSO4.
For FtsANg-FtsNNg and FtsANg-FtsQNg SPR experiments, ligands (i.e. His-FtsNNg for FtsANg-FtsNNg, and His-FtsQNg for FtsANg-FtsQNg interactions) were immobilized onto the sensor chip at a concentration of 200 nM. A two-fold dilution series of the analyte (FtsANg), in PBS buffer with Tween-20 (PBST; 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, 0.1% BSA, 0.05% Tween-20, pH 7.9), was injected at a flow rate of 30 μl/min over the surface of the chip for 120 s. This was followed by an injection of PBST buffer for 300 s. Negative controls comprised a reference channel flowed with PBST buffer, and a chip surface immobilized with either FtsQNg or FtsNNg flowed with GST in PBST.
For the FtsANg-FtsZNg interaction, the SPR binding assay was performed using MES buffer, supplemented with 0.1% BSA, 0.05% Tween-20, and 1 mM ATP were added with the pH adjusted to 7.5. FtsANg was immobilized on the chip surface as described above. Each 120-s injection of polymerized FtsZNg was followed by an injection of supplemented MES buffer for 300 s for dissociation. Negative controls included a reference channel which was flowed with MES buffer containing 2 mM GTP, and the FtsANg-immobilized chip surface flowed with GST in supplemented MES instead of polymerized FtsZNg.
All SPR data was analyzed using ProteOn Manager™ (Bio-Rad Laboratories). The sensorgram (i.e. a graph of the response unit versus time) was first subtracted by the response units (RU) of the reference channel, with no immobilized ligands, to reduce the non-specific binding signals between analyte and empty chip surface. Then, the sensorgram was subtracted with the RU signal with running buffer and ligand immobilized on the chip. Association and disassociation constants were obtained using the Langmuir 1:1 kinetic fit model, by nonlinear regression, using ProteOn Manager™. Each protein pair was tested minimally in duplicate.
Using B2H assays, we investigated 28 potential interactions among eight gonococcal divisome proteins including FtsZ, FtsA, ZipA, FtsK, FtsQ, FtsI, FtsW, and FtsN. The results (Table 4) show that nine interactions, FtsZ-FtsA, FtsZ-FtsK, FtsZ-FtsW, FtsA-FtsK, FtsA-FtsQ, FtsA-FtsW, FtsA-FtsN, FtsI-FtsW, and FtsK-FtsN, displayed a residual β-galactosidase activity lower than 50%, indicating a positive interaction between these proteins in N. gonorrhoeae. The interaction between FtsANg and FtsNNg had the lowest residual β-galactosidase activity (24%), indicating the strongest interaction. This was followed by FtsANg-FtsKNg (30%), FtsNNg-FtsKNg (31%), FtsINg-FtsWNg (35%), FtsZNg-FtsWNg (39%), FtsANg-FtsZNg (40%), FtsZNg-FtsKNg (41%), FtsANg-FtsWNg (45%), and FtsANg-FtsQNg (48%) interactions. ZipANg did not directly interact with other cell division proteins as the residual β-galactosidase activity of all interactions was above 50% (Table 4).
To confirm the results of selected B2H assays, we examined several interactions (i.e. FtsQNg-FtsANg, FtsANg-FtsNNg, FtsANg-FtsZNg) using GST pull-down assays. GST pull-down results (Fig. 1a) showed that His-FtsQNg was pulled down by GST-FtsANg, but not GST itself (negative control), indicating an interaction between FtsANg and FtsQNg. Using similar evaluation criteria, we ascertained that His-FtsANg was pulled down by GST-FtsNNg, indicating an interaction between these two proteins (Fig. (Fig.1b1b).
The interactions of FtsANg and FtsZNg from E. coli in vitro requires the presence of both ATP and GTP . GTP promotes FtsZ polymerization, and ATP is necessary for FtsA to interact with FtsZ, but not for FtsZ polymerization [36, 37]. The presence of FtsZNg polymers in MES buffer was determined by transmission electron microscopy (TEM) and dynamic light scattering (DLS; Additional file 2: Figure S2). GST pull-down assay did not detect an interaction between FtsANg and FtsZNg in the presence of 1 mM ATP and 2 mM GTP (Fig. (Fig.1c).1c). This result was unexpected, given our B2H results and the commonality of FtsA-FtsZ interaction in other bacterial species [24, 25, 38, 39], as ascertained by different in vivo assays (i.e. B2H, yeast two-hybrid, chemical cross-linking with co-immunoprecipitation).
Surface plasmon resonance (SPR) was used to confirm selected gonococcal cell division protein-protein interactions in real-time. SPR was used to evaluate the interactions of FtsANg with FtsZNg because of the conflicting results observed with B2H and GST pull-down assays. GTP was added to promote FtsZNg polymerization (Additional file 2: Figure S2). The sensorgram indicated that FtsZNg interacted with FtsANg at concentrations of 6 μM and 12 μM (Fig. 2a), but not at concentrations lower than 6 μM (data not shown). Kinetic analysis showed that the FtsANg-FtsZNg interaction had a slow association (ka=3.56×102 M−1 s−1) and a significant disassociation activity (kd=5.31×10−3 s−1), giving a KD value of 14.9 μM. This suggested that the interaction between FtsANg and FtsZNg was likely transient. When GTP was absent from the FtsZNg protein solution, no binding was detected between FtsANg and FtsZNg (data not shown). The sensorgram of the interaction between FtsANg and the negative control (GST) also showed no binding activity (Fig. (Fig.2b),2b), indicating the specificity of the SPR results for the interaction of FtsANg with FtsZNg.
For the SPR analysis of the FtsANg-FtsQNg interaction, FtsANg was tested using various concentrations (from 31.25 nM to 250 nM; Fig. Fig.2c).2c). At 0 s, the association of FtsANg and FtsQNg was observed immediately following injection of the FtsANg solution onto the FtsQNg-labeled chip surface, with a rapid increase of response units (ka=2.72×105 M−1 s−1; Fig. Fig.2c).2c). This indicated a fast binding event between the two proteins. Disassociation between FtsANg and FtsQNg was not significant (kd=4.09×10−3 s−1), suggesting this interaction was strong and stable (KD=15.1 nM). The negative control, using non-interacting GST, did not cause any change in the response units (Fig. (Fig.2d2d).
The FtsANg-FtsNNg interaction was observed with an increasing concentration of FtsANg (62.5 nM, 125 nM, 250 nM and 500 nM; Fig. Fig.2e).2e). His-FtsNNg had a binding affinity (KD) of 53.3 nM with FtsANg. The association and disassociation constants were 1.15×105 M−1 s−1, and 6.16×10−3 s−1, respectively (Fig. (Fig.2e),2e), indicating a strong interaction between FtsANg and FtsNNg. The injection of non-interacting GST onto the FtsNNg immobilized chip surface did not cause any change in the response units (Fig. (Fig.2f2f).
Since FtsANg interacted with FtsZNg, FtsQNg, FtsWNg, and FtsNNg, we further examined the interaction regions of FtsANg with these four proteins using B2H assays. Based on FtsANg homology modeling, six FtsANg truncations (T1-T6) were created (Additional file 1: Figure S1), which contained one or more FtsANg subdomains . FtsZNg self-interaction was used as a positive control. And negative controls included E. coli R721 without plasmids or carrying each single recombinant B2H vector in which the gene of interest had been cloned. FtsANg truncations T3, T4, and T5 interacted with FtsZNg and FtsNNg (Figs. 3 and and4,4, blue bars). FtsANg truncations T1, T2, and T6 did not show an interaction with these proteins (Figs. (Figs.33 and and4,4, green bars). The T4 and T5 truncations included the 2B and 2A2 subdomains of FtsANg, suggesting that these subdomains of FtsANg interacted with both FtsZNg and FtsNNg. The T3 construct contained also contained the 2A1 subdomain of FtsANg, as compared to truncations T1 and T2, indicating that this subdomain was also involved in interactions with FtsZNg and FtsNNg. FtsQNg interacted only with the T4 and T5 truncations of FtsANg (Fig. 5, blue bars), indicating that the 2B and 2A2 subdomains, but not the 2A1 subdomain, were required for the FtsANg-FtsQNg interaction. Only the T5 truncation of FtsANg interacted with FtsWNg, suggesting that 2A2 subdomain was involved in the interaction with FtsWNg (Additional file 3: Figure S3). In summary, these results showed that the 2A1, 2A2 and 2B subdomains of FtsANg are required for its interaction with FtsNNg and FtsZNg. The FtsANg 2A2 and 2B subdomains are required for interaction with FtsQNg, and the 2A2 subdomain is involved in the interaction with FtsWNg.
The N. gonorrhoeae cell division interactome described in our study is the third cell division interaction network identified in bacteria, in addition to E. coli and S. pneumoniae (Fig. 6a) [23–25]. Compared to the other two interactomes (Fig. 6b and c), fewer interaction protein pairs are identified in N. gonorrhoeae (Fig. (Fig.6a).6a). Only nine interactions are present among the eight divisome proteins tested in N. gonorrhoeae, while E. coli and S. pneumoniae have 21 and 17 interactions among ten and eight divisome proteins, respectively [24, 25].
The development of all three cell division interactomes was based on interaction data obtained from the same B2H system [24, 25] The E. coli interactome was developed using B2H results exclusively while the S. pneumoniae study also applied co-immunoprecipitation to verify selected B2H positive interaction pairs [24, 25]. In our study, we used a combination of GST pull-down and surface plasmon resonance to further study selected positive B2H interactions.
Two interactions, FtsA-FtsZ and FtsZ-FtsK, are conserved in the cell division interactomes of N. gonorrhoeae, E. coli and S. pneumoniae (Fig. (Fig.6,6, red lines). The FtsA-FtsZ interaction is a common interaction in prokaryotes [24, 25, 39–41]. Both our B2H and SPR results confirmed this interaction in N. gonorrhoeae. A proper ratio between FtsA and FtsZ is crucial for the interaction in E. coli  and our SPR results support this finding; FtsANg interacts with FtsZNg only when its concentration is higher than 6 μM (Fig. 3b), indicating that the interaction requires a critical concentration threshold. Our SPR results further showed that interaction between FtsANg and FtsZNg was transient, a result warranting further study to fully understand its implications for divisome formation in N. gonorrhoeae. Unexpectedly, the GST pull-down assay, an in vitro assay, did not detect an FtsANg-FtsZNg interaction. We believe that this “false negative” in vitro result was caused by the requirement of a membrane/solid surface support for the interaction to anchor FtsA [35, 43, 44].
The interaction of FtsZ with FtsK has been observed in N. gonorrhoeae, E. coli, S. pneumoniae, B. subtilis and C. crescentus [24, 25, 45, 46]. The C-terminus of FtsK is required for proper DNA segregation in E. coli . The absence of an FtsZ-FtsK interaction in both E. coli and C. crescentus caused abnormal chromosome segregation and cell filamentation [45, 48]. This suggests that the FtsZ-FtsK interaction connects the cell division process with chromosome segregation, by ensuring that the replicated chromosome is cleared from the division site.
The FtsA-FtsW interaction has been observed only in N. gonorrhoeae (Fig. (Fig.6,6, blue lines). Since FtsW is a membrane protein and difficult to purify, we did not verify the interaction by GST pull-down and SPR assays. However, we performed additional B2H assays to identify which subdomains of FtsA were involved in its interaction with FtsW (Additional file 3: Figure S3) and showed that the 2A2 subdomain of FtsA strongly interacts with FtsW (Additional file 3: Figure S3). FtsW, an inner membrane protein, is required in E. coli for the recruitment of FtsI and the translocation of the cell well precursor, lipid II [20, 21, 49, 50]. An FtsI-FtsW protein interaction has been observed in E. coli, Streptomyces coelicolor, and Mycobacterium tuberculosis [21, 51, 52]. Interestingly, we discovered that FtsINg only interacts with FtsWNg, suggesting that its localization may depend on this protein.
The importance of the unique FtsKNg-FtsNNg interaction in N. gonorrhoeae, as determined by B2H, is not clear (Fig. (Fig.6,6, blue lines). In E. coli, FtsN is the last protein, of ten essential cell division proteins, recruited to the division site to initiate cell constriction [53, 54]. A previous study suggested that E. coli FtsN and FtsK stabilize the Z-ring cooperatively, without direct interactions . Since the FtsK-FtsN interaction is present in N. gonorrhoeae, their joint involvement in gonococcal cell division requires further investigation.
ZipANg did not interact with any other gonococcal cell division protein. In E. coli, ZipA only interacts with FtsZ, and is required for downstream protein recruitment, including FtsK, FtsQ, FtsL, and FtsN [24, 56]. One report suggested that ZipANg is a homologue of the E. coli protein with high similarity in its key domains . Although ZipANg complemented a conditional zipA mutant in E. coli, it did not fully restore a wild type phenotype in this strain . Given these data, the role of ZipA in gonococcal cell division remains to be elucidated.
In N. gonorrhoeae, the existence of FtsLNg is unclear due to its low homology with E. coli FtsL . An open reading frame (ORF) located between mraW and ftsI in the dcw cluster of N. gonorrhoeae was reported by Francis et al.  and they reported that it was not a coding ORF. Snyder et al.  named the same ORF ftsL. Because this ORF shares only 17% amino acid similarity to its E. coli homologue, we considered that it was not a functional ORF and did not test its interaction with other gonococcal cell division proteins.
N. gonorrhoeae lacks FtsB ; thus, the protein complex FtsQ-B-L, present in other species, such as E. coli, S. pneumoniae and B. subtilis, would not be formed in N. gonorrhoeae [59–61]. This protein complex has been described as a bridge connecting FtsK and the FtsI-FtsW complex in E. coli . A recent study suggests that the E. coli FtsQ-B-L complex acts as a signal transmitter for cell wall remodeling and constriction, which is mediated by direct interactions with the FtsI-W complex and FtsN . In S. pneumoniae, the FtsQ homologue, DivIB, interacts with FtsKSp, FtsLSp, and FtsWSp . Interestingly, our B2H data show that FtsQNg only interacts with FtsANg, suggesting that the function of FtsQNg in cell division in N. gonorrhoeae may be distinct.
There are several models for bacterial cell constriction. One E. coli model suggests that the force that drives constriction comes from septal peptidoglycan synthesis . In this model, the FtsAEc-FtsNEc interaction activates peptidoglycan synthesis by direct or indirect interaction with FtsIEc . Another E. coli model suggests that the energy generated from FtsZ-mediated GTP hydrolysis drives cell constriction . We observed an FtsANg-FtsNNg interaction in N. gonorrhoeae. However, there is no further evidence supporting either model of cell constriction in N. gonorrhoeae at this time.
The non-essential proteins, FtsENg and FtsXNg, are also implicated in cell division in N. gonorrhoeae . Similarly, in E. coli, FtsE and FtsX are non-essential for cell division under conditions of high osmotic pressure . Gonococcal FtsE and FtsX have high similarity in amino acid sequence to known homologues in other species . In E. coli, the interaction between FtsE and FtsZ has a regulatory effect on the Z-ring . Future research could focus on revealing the effects of FtsENg and FtsXNg on cell division in N. gonorrhoeae.
The major issue interpreting B2H assay results is the empirical cut-off of 50% residual ß-galactosidase activity used to discriminate positive and negative interactions. In particular, values close to the cut-off could be interpreted as either false positive or negative results. To validate our B2H results, we used other B2H interactions to test which subdomains of FtsANg interacted with FtsZNg, FtsNNg, and FtsQNg. We determined that the 2A and 2B subdomains of FtsANg interacted with FtsZNg, FtsQNg, and FtsNNg. We also evaluated some positive interactions obtained by B2H using SPR and GST pull-down assays. The SPR method detects and measures weak or transient interactions, in real-time, with high sensitivity . The SPR method showed a transient FtsANg-FtsZNg interaction. GST pull-down assays, on the other hand, are ideal in detecting strong protein-protein interactions, as weak interactions may dissociate during the assay . We consider this to be a reasonable explanation for our failure to confirm when the interaction of FtsANg with FtsZNg when using a GST pull-down assay.
To date, most of studies on cell division have been focused on model organisms (i.e. the Gram-negative rod E. coli and the Gram-positive rod B. subtilis) due to the abundant availability of tools for genetic manipulation . Research on cell division in non-model organisms is expanding, and this includes studies with N. gonorrhoeae [7, 27]. For example, Chlamydia trachomatis, which lacks FtsZ, requires an actin-like protein, MreB, for cell division . A gene cluster encoding three cell division proteins, named MldA, MldB, and MldC, was identified only in Clostridium difficile and its closely related bacteria . Results from studies using non-model organisms suggest that cell division mechanisms are complex and vary in different organisms, reflecting vast biological diversity.
In our research, we discovered that nine interactions among eight cell division proteins defined the cell division interactome of N. gonorrhoeae. In comparison with the published cell division interactomes of E. coli and S. pneumoniae, FtsA-FtsZ and FtsZ-FtsK interactions were common to all three bacteria. FtsK-FtsN and FtsA-FtsW interactions were only present in N. gonorrhoeae, suggesting that they play different roles in the cell division of this microorganism. ZipANg did not interact with any other cell division proteins tested in this study, indicating that its role may differ as compared to its E. coli homologue. We also determined that the subdomains of FtsANg which interacted with FtsQNg, FtsZNg, FtsWNg, or FtsNNg, differed from its E. coli homologue. This suggests that N. gonorrhoeae possesses a distinctive cell division interactome, and likely a different mechanism of cell division as compared to E. coli and other organisms.
Schematic representation of N. gonorrhoeae ftsA and its truncations . T1 (162aa, Met1-Ala162) contained the N-terminal 1A and 1C domains of ftsA Ng. T2 (194aa, Met1-Val194) included the N-terminal 1A, 1C and 1A domains of ftsA Ng. T3 (230aa, Met1-Ile230) included the N-terminal 1A, 1C, 1A and 2A1 domains of ftsA Ng. T4 (71aa, Pro231-Glu301) contained the 2B domain of ftsA Ng. T5 (114aa, Ile301-Leu414) contained the 2A2 and 1A C-terminal domains of ftsA Ng. T6 (64aa, Ala351-Leu414) contained the 1A C-terminal domain of ftsA Ng. (DOCX 30 kb)
FtsZNg polymerization assays. FtsZNg polymers visualized by transmission electron microscope with(A) or without (B) 2 mM GTP in MES buffer (50 mM MES-NaOH, 50 mM KCl, 10 mM MgCl2, pH 7.5) at 30 °C. Solid arrows indicate FtsZNg polymers. Scale bar indicates 100 nm. (C) Light scattering of FtsZNg polymerization (6 μM) in MES buffer. (DOCX 211 kb)
Interactions between FtsANg truncations (T2, T3, T4, T5 and T6) and FtsWNg (W) by B2H assay. Values of less than 50% (<1250 Miller Units) indicate a positive interaction between two proteins (blue bars) while values of more than 50% (>1250 Miller Units) indicate a negative interaction (green bars). (DOCX 66 kb)
The authors wish to give special thanks to Jason Maley from the Saskatchewan Structural Sciences Centre, University of Saskatchewan for help with surface plasmon resonance and dynamic light scattering experiments. Electron microscopy was performed at the Western College of Veterinary Medicine, University of Saskatchewan.
This work was supported by the Saskatchewan Health Research Foundation New Investigator Establishment Grant (Grant #1866–2007 to JRD) as well as the Natural Sciences and Engineering Research Council of Canada Grant (Grant #203651–2012 RGPIN to JRD). Yan Li and Yinan Zou were partially supported by graduate scholarships from the University of Saskatchewan.
The data from this report are included within the article. Datasets used in the current study are available upon request.
YZ and YL participated in the experimental design, implementation and data analysis. YZ also wrote the first draft of manuscript. JRD designed and supervised the entire project and was responsible for the final submission of the manuscript. All authors contributed to manuscript revisions. All authors have read and approved the final manuscript.
Ethics approval is not required for this study. Consent to participate is not applicable.
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
The online version of this article (10.1186/s12866-017-1140-1) contains supplementary material, which is available to authorized users.
Yinan Zou, Email: firstname.lastname@example.org.
Yan Li, Email: moc.liamg@9011aloul.
Jo-Anne R. Dillon, Phone: 1 306 9664232, Email: email@example.com.