Interactions between the Aggregatibacter actinomycetemcomitans secretin HofQ and host cytokines indicate a link between natural competence and interleukin-8 uptake

Figures & data

Figure 1. emHofQ Interacts with Cytokines. (A) Binding of cytokines by emHofQ was determined with ELISA. Data are presented as percentage of binding to control protein BSA. Binding of emHofQ to IL-8 was significantly higher than that to BSA (p-value 0.029, Mann-Whitney U-test); N = 4. (B) Interaction with IL-8 and IL-1β was measured with varying emHofQ concentrations using TRFIA. Dissociation constants were calculated for the binding curves: 43 ± 4 nM (emHofQ-IL-8) and 140 ± 20 nM (emHofQ-IL-1β). Data were obtained from three independent experiments and are presented as the mean± SD. BSA was used as a negative binding control. (C) Binding of IL-8 to immobilized emHofQ at different IL-8 concentrations was determined by surface plasmon resonance. The dissociation constant was estimated from the binding curves as being 2.4 ± 1.3 μM. Three independent experiments were performed; one representative image is shown as the primary figure, and the two other images are shown as insets. (D) In the thermal shift assay, the melting temperature (T_m) of emHofQ (45.3°C) increased when emHofQ and IL-8 were co-incubated. At the highest tested IL-8 concentration of 200 μM, the T_m obtained was 51.7°C; N = 4–6. (E) In contrast, the T_m of IL-1β (60.2°C) decreased when IL-1β was incubated with emHofQ. With 50 μM emHofQ, the T_m obtained was 55.8°C; N = 3–4.

Figure 2. The type I KH domain of emHofQ interacted with IL-8. (A) Homodimeric structure of emHofQ (PDB: 2Y3M). Monomer A is colored blue-red from the N terminus to the C terminus, and monomer B is in gray. K139, which involved in the IL-8 interaction, and as the secondary structure elements β5 and α3, which are located next to the interacting loop, are indicated. (B) A total of five mutations were made in two loops (L1 and L2) in the type I KH domain of emHofQ: L1 (H136A, F137A, K139A) and L2 (D165A, R166A). (C) Mutated emHofQ variants bound to IL-8 and IL-1β with lower affinity than the wild-type emHofQ. BSA was used as a negative binding control (smaller graph). Data were obtained from three independent experiments and are presented as the mean± SD. Dissociation constants (K_d) and values of maximum binding (B_max) for each variant are listed in the table. (D) Structural comparison of the C-terminal type I KH domain of emHofQ (residues 126–191, in gray) with the N1-domain of NmPilQ (residues 439–516 (PDB id: 4AV2), in blue).

Figure 3. Recombinant IL-8 interacted with linear dsDNA. Linear A. actinomycetemcomitans USS DNA (31 μg/ml) was incubated with varying amounts of IL-8 for 30 min at RT and run on a 0.8% agarose gel (EMSA). Interaction between the DNA and IL-8 was observed with high IL-8 amounts as the formed complexes were trapped in the wells. BSA (10 mg/ml) was used as a negative control.

Figure 4. Verification of the ΔhofQ mutant and expression of hofQ from the com locus. (A) Deletion of the hofQ gene from A. actinomycetemcomitans D7S was verified by extracting the genomic DNA and performing PCR with flanking primers. The deletion was further verified at the transcriptional level by extracting RNA and performing reverse transcription -PCR. No hofQ gDNA or cDNA was observed in the A. actinomycetemcomitans ΔhofQ mutant. In addition, in the D7S wild-type strain, hofQ was transcribed at a higher level in minimal (RPMI) medium than in rich (TSB) culture medium. The control gene clpX was expressed and transcribed at the same level in both strains and culture conditions. (B) Genetic organization of the com locus in A. actinomycetemcomitans D7S. DNA in the com locus is marked with dark arrows, and additional ORFs outside the com locus are noted with gray arrows. Small black arrows indicate the locations of the primers (numbered) used for operon characterization. (C) Neither the gene locus from comA to shiKi nor that from comA to hofQ was detected as a single transcription unit, even in wild-type D7S, indicating that these genes are organized in different operons. However, the comD and hofQ genes were co-transcribed in wild-type D7S but were absent in the ΔhofQ strain. In wild-type A. actinomycetemcomitans, the comD-hofQ operon was transcribed more in minimal culture medium (RPMI) than in rich culture medium. Primers used for operon characterization are indicated by their names and numbers (listed in Table 4) next to each panel.

Figure 5. Deletion of hofQ increased the antimicrobial susceptibility of A. actinomycetemcomitans to β-lactams. Deletion of the hofQ gene made A. actinomycetemcomitans more susceptible to ampicillin (amp, p = 0.001) and amoxicillin/clavulanic acid (amc, p = 0.001) but did not significantly alter the susceptibility to tetracycline (tet, p = 0.093) or doxycycline (dox, p = 0.485). MIC values were determined from the scale on the Etest® strip after 45–48 h of incubation. Data were obtained from 5–9 independent experiments and are presented as the mean± SD.

Figure 6. emHofQ was not antigenic in A. actinomycetemcomitans-positive periodontitis patients, and HofQ did not enhance the production of ROS by leukocytes when stimulated with A. actinomycetemcomitans cells. (A) Compared with the wild-type strain, the hofQ deletion did not significantly affect the production of ROS by leukocytes (p = 0.442). ROS production was measured as chemiluminescence, and the peak values obtained for the wild-type and ΔhofQ mutant A. actinomycetemcomitans strains were compared. Bacteria were pretreated with plasma collected from a healthy individual, and leukocytes were freshly isolated from the plasma of healthy control subjects (N = 8). (B) Opsonization efficiency of the sera collected from A. actinomycetemcomitans-positive patients was higher than that of the healthy control sera, but these sera significantly only enhanced the production of ROS by leukocytes when the ΔhofQ mutant was used (p = 0.005). Peak values of the chemiluminescence reactions are reported for 12 patient sera and 12 control sera. The horizontal lines indicate the mean values of each sample set; N = 1–3. (C) Sera collected from A. actinomycetemcomitans-positive patients did not have higher emHofQ antibody levels than sera from healthy control subjects when tested with ELISA. A total of 12 patient sera and 12 control sera were tested. Data corresponding to nonspecific binding to the negative control (BSA) was subtracted from the data obtained for emHofQ binding.

Table 1. Formation and composition of A. actinomycetemcomitans D7S ΔhofQ biofilm compared to that of the wild-type (wt) D7S strain.

	Biofilm mass	eDNA	PGA	Total protein
% of wt	45 ± 22	57 ± 12	78 ± 9	54 ± 13
(p-value)	(0.029)	(0.029)	(0.029)	(0.029)

Table 2. Effect of hofQ deletion on the biofilm response to cytokines IL-1β and IL-8, measured as changes in biofilm composition.

		Percentage of the control (without cytokines)
Strain	Cytokine	Biofilm mass	eDNA	PGA	Total protein
D7S wt	IL-1β	103 ± 7	70 ± 14	105 ± 5	85 ± 11
p-values			(0.029)	(0.029)	(0.029)
D7S wt	IL-8	96 ± 2	65 ± 18	88 ± 19	90 ± 20
p-values		(0.029)	(0.029)
D7S ΔhofQ	IL-1β	95 ± 15	116 ± 29	98 ± 5	102 ± 8
D7S ΔhofQ	IL-8	89 ± 15	113 ± 39	97 ± 7	100 ± 29

Table 3. Codon-optimized genes for protein expression. Restriction enzyme binding sites (NdeI and XhoI) are underlined.

Gene name	Codon-optimized sequence
mature_il-1β	CAT ATG GCC CCT GTA CGT TCC CTC AAT TGC ACC TTA CGC GAT AGT CAG CAG AAA TCT TTG GTG ATG TCA GGT CCC TAC GAA CTG AAA GCA CTC CAT CTG CAA GGA CAG GAT ATG GAA CAG CAA GTC GTG TTT AGC ATG AGC TTT GTA CAA GGC GAA GAA TCG AAC GAC AAA ATT CCA GTT GCT CTT GGC CTG AAA GAA AAG AAC CTG TAT CTG TCG TGT GTC TTG AAA GAT GAT AAA CCG ACA CTG CAG TTA GAA AGT GTT GAT CCG AAG AAT TAT CCG AAG AAG AAA ATG GAG AAA CGC TTT GTG TTC AAC AAA ATC GAG ATC AAC AAC AAA CTG GAG TTC GAA TCT GCG CAA TTT CCG AAT TGG TAC ATT AGC ACC TCA CAG GCG GAA AAT ATG CCA GTG TTT CTG GGT GGG ACT AAA GGC GGT CAG GAC ATT ACG GAC TTC ACC ATG CAG TTC GTT TCC AGC TAA CTC GAG
emhofQ_L1	CAT ATG CAG AAT CCG GTG TTC TCG ATT CGC TTG AAA CAA GCG CCA CTC GTA CCG ACT TTA CAG CAG TTG GCG TTA GCG CAT AAC ACC AAC CTG ATT ATC GAC GAT GAA CTG CAA GGT ACG GTG AGT CTG CAA CTG GAA AAC GTG GAT CTG GAC CAG CTC TTT CGT AGC GTT GCG AAG ATT AAG CAG CTG GAT CTG TGG CAG GAG AAT GGC ATC TAT TAC TTC ACC AAA GGC GAT ACC AAC ACG AAA TTT GCG GGG AAA ATG GAG GAA CCG TTT CCG TTA TCT CTG CCT ATG GCC GAA GAA CCC GCT CAG CTT AAC ACA GCC ACC ATC AAA CTT GCA GCT GCA GCA GCC TCT GAA GTC ATG AAA TCG CTC ACT GGT GGT TCA GGA AGC CTG CTT TCC CCA AAT GGC TCC ATT ACG TTC GAT GAC CGC AGC AAT TTG CTG CTG ATT CAG GAT GAA CCT CGT AGT GTT CGC AAC ATC AAG AAA CTG ATC AAA GAG CTG GAC AAA CCG ATT GAA CAA CTC GAG
emhofQ_L2	CAT ATG CAG AAT CCG GTG TTC TCC ATT CGT CTG AAA CAA GCA CCT CTC GTT CCG ACA CTT CAG CAG TTA GCG TTA GCC CAT AAC ACC AAC CTG ATT ATT GAC GAT GAA CTC CAA GGC ACG GTT AGC CTG CAA CTG GAG AAT GTG GAC TTG GAT CAG CTG TTT CGC AGT GTA GCG AAA ATC AAA CAG CTG GAT CTT TGG CAG GAA AAC GGG ATC TAC TAT TTC ACC AAA GGT GAT ACG AAC ACC AAA TTT GCG GGC AAA ATG GAG GAA CCG TTT CCG CTT AGC CTG CCA ATG GCA GAA GAA CCA GCT CAG TTG AAT ACT GCG ACG ATC AAA CTG CAC TTT GCC AAA GCT TCC GAA GTG ATG AAG TCG TTG ACT GGT GGT AGC GGA TCT CTG CTC TCA CCG AAT GGC TCG ATT ACC TTC GAT GCA GCC TCT AAC CTG TTA CTG ATT CAA GAC GAA CCT CGT AGT GTC CGC AAC ATC AAG AAG CTG ATC AAA GAG CTG GAT AAA CCC ATT GAA CAG CTC GAG
emhofQ_L1L2	CAT ATG CAG AAT CCC GTT TTC AGC ATT CGC CTG AAA CAA GCG CCT TTG GTA CCG ACC CTT CAG CAG TTA GCC TTG GCG CAT AAC ACC AAC CTG ATT ATT GAC GAT GAA CTG CAA GGT ACT GTG TCT CTG CAG TTA GAG AAT GTG GAT CTG GAC CAG TTG TTC CGT TCA GTC GCG AAA ATC AAA CAG CTT GAT CTG TGG CAA GAG AAC GGC ATC TAC TAT TTC ACG AAG GGT GAT ACC AAC ACC AAA TTT GCC GGG AAA ATG GAA GAA CCG TTT CCG CTG TCG TTA CCT ATG GCT GAA GAA CCA GCG CAA CTC AAC ACA GCG ACG ATC AAA CTG GCT GCT GCA GCC GCA AGT GAA GTG ATG AAA TCC CTG ACT GGA GGT AGT GGC TCT CTG CTG TCC CCG AAT GGC TCG ATT ACG TTT GAT GCC GCA AGC AAT CTT CTC CTG ATT CAG GAT GAA CCA CGT AGC GTT CGC AAC ATC AAG AAG CTC ATC AAA GAG CTG GAC AAA CCG ATT GAA CAG CTC GAG

Table 4. Primers and annealing temperatures used in this study. Restriction enzyme sites are underlined.

D7S ΔhofQ generation
Primer name (code)	Sequence 5`→ 3`	Tm annealing (°C)
HofQ_nest_F (01)	TTGGTTCCGTTCCTTTAATGAAATATTC	70.0
HofQ_nest_R (02)	ATAGCATGACCAAAGGTATGACCGAGA
ComB_F (03)	TACAGCCACAGTCAAGAAGAA	60.6
HofQ_R_BamHI (04)	TATGGATCCTGCTATTTCCTTCAAAAGTTC
HofQ_F_SalI (05)	TAAGTCGACCTAAATAGGAAATAA	46.2
3DHQ synthase_R (06)	GTAGCTTCTTTCTTTCAA
Spe_Urf (07)	GCCACTGCATTTCCCGCATA	69.1
Spe_DWR2 (08)	TGCAGGTCGATTTTCGTTCGT
HofQ_seq_F1 (09)	ATGTTACACCCTCTTTTC	51.2
HofQ_seq_R1 (10)	TGCCATTGTTATAATCTTA
HofQ_F1_seq2	ATTCTATTCACAGCGAAC	*
3DHQ_R2_seq2	TTCAGCGGATAACAATG
Reverse transcription PCR
16 S_F	TACGCATTTCACCGCTACA	59.1
16 S _R	ATGCCAACTTGACGTTAAAT
HofQ_F_rt1 (13)	TCTTATCGTCAGCCAAAACTCG	63.9
HofQ_R_rt1 (14)”	TATCGCTAAATACACCGCCC
ClpX_F (15)	GGTTATGTGGGCGAAGATG	64.0
ClpX_R (16)	CGATTTGCGGGTGATTTTG
ComA_F (17)	TTGCCCCATAACCTGAAC	57.8
ShiKi_I_R (18)	GTAGATAACACGATGCCTT
ComD_F (19)	TTTGCGTTGGGATAACCC	63.7
HofQ_R (20)	ACAGACCGCACTTTAGCCAT

*performed by Eurofins Genomics

”Annealing temperature of 61.5°C with primer pair 14/17

Figure 7. Preparation of the markerless D7S ΔhofQ mutant. Natural transformation and homologous recombination was used to replace hofQ with the recombinant loxP-spe-loxP cassette that contained the flanking regions upstream and downstream of hofQ. The transformants (ΔhofQ::spe) were verified by PCR. The loxP-spe-loxP cassette was removed by the pAT-Cre-plasmid (Tet^R), which was introduced via natural transformation. Cre recombinase then catalyzed the recombination between the loxP sites. The obtained colonies were examined for Spe and Tet resistance, and the colonies sensitive to both antimicrobials had lost both the loxP-spe-loxP cassette and the pAT-Cre plasmid, i.e., these colonies were markerless D7S ΔhofQ mutants.