Substitutions in the nonactive site of the passenger domain on the activity of Haemophilus influenzae immunoglobulin A1 protease

ABSTRACT

Immunoglobulin A1 (IgA1) protease is a critical virulence factor of Haemophilus influenzae that facilitates bacterial mucosal infection. This study investigates the effect of iga gene polymorphism on the enzymatic activity of H. influenzae IgA1 protease. The IgA1 protease activity was examined in the H. influenzae Rd KW20 strain and 51 isolates. Genetic variations in iga and deduced amino acid substitutions affecting IgA1 protease activity were assessed. Machine learning tools and functional complementation assays were used to analyze the effects of identified substitutions on the stability and activity of IgA1 protease, respectively. All 51 isolates exhibited similar iga expression levels. No igaB expression was detected. According to comparisons with the reference Rd KW20 strain, four substitutions in the protease domain, 26 in the nonprotease passenger domain, and two in the β-barrel domain were associated with the change in IgA1 protease activity. No substitutions in the catalytic site of IgA1 protease were observed. Logistic regression, receiver operating characteristic curves, Venn diagrams, and protein stability analyses revealed that the substitutions Asn352Lys, Pro353Ala, Lys356Asn, Gln916Lys, and Gly917Ser, which were located in the nonactive site of the passenger domain, were associated with decreases in IgA1 protease activity and stability, whereas Asn914Lys was associated with an increase in these events. Functional complementation assays revealed that the Asn914Lys substitution increased IgA1 protease activity in the Rd KW20 strain. This study identified substitutions in the nonactive site of the passenger domain that affect both the activity and stability of H. influenzae IgA1 protease.

INTRODUCTION

Bacterial immunoglobulin A1 (IgA1) proteases are proteolytic enzymes that cleave specific proline-serine and proline-threonine peptide bonds in the hinge region of human IgA1 (1, 2). Because IgA1 is critical to immune defense, particularly at mucosal surfaces, IgA1 proteases secreted by pathogenic microbes destroy host IgA1, thereby enervating host defense at mucosal sites of infection. IgA1 proteases have been identified in Haemophilus influenzae, Streptococcus pneumoniae, Neisseria meningitidis, and Neisseria gonorrhoeae and are crucial virulence factors for these bacterial species (3 – 6). In H. influenzae, IgA1 protease activity is significantly higher in strains isolated from sputum, blood, cerebrospinal fluid, or sterile tissue of symptomatic individuals than in those isolated from throat swabs of asymptomatic carriers (7). Furthermore, the upregulation of H. influenzae IgA1 protease was observed in a model of experimental human nasopharyngeal colonization (8). These findings reveal the pathogenic effects of IgA1 protease on infection, invasion, and early colonization of H. influenzae in the host.

H. influenzae IgA1 protease is encoded by the iga (HI_0990) gene and functions as an autotransporter consisting of an N-terminal signal sequence, an internal passenger domain containing a protease domain, and a C-terminal β-barrel domain. The iga gene from the H. influenzae prototypic strain Rd contains 5,082 base pairs and encodes a protein with a molecular mass greater than 180 kilodaltons. In addition to the common iga gene, a second IgA1 protease gene, igaB, was characterized in nontypeable H. influenzae by Fernaays et al. (9). This gene is located within the region between HI_0184 and HI_0164 and exhibits a high sequence homology to the iga of N. meningitidis and N. gonorrhoeae. Later on, Murphy et al. reported the expression of the igaB gene in approximately 40% of the H. influenzae strains isolated from the respiratory tract of adult patients with chronic obstructive pulmonary disease (COPD) and identified the distribution of IgA protease variants, including IgA-A1, IgA-A2, IgA-B1, and IgA-B2, which cleave different sites at the hinge region of IgA1 (10). Furthermore, a high detection rate of igaA1 was observed to be associated with the exacerbation of H. influenzae infection (10). These insights suggest a correlation between the expression pattern of IgA1 proteases and the severity of H. influenzae diseases. However, some H. influenzae strains isolated exhibited a weak or loss of IgA1 protease activity. In the present study, we assessed the relevance of gene polymorphism in the phenotype of H. influenzae IgA1 protease and investigated the effects of amino acid substitutions on the stability and activity of H. influenzae IgA1 protease.

MATERIALS AND METHODS

Bacterial isolates and sequencing

This study was approved by the E-Da Hospital Institutional Review Board (No. 2023025) in accordance with the ethical standards noted in the 1964 Declaration of Helsinki and its later amendments. The collection, culture, preservation, and whole-genome sequencing of H. influenzae isolates have been described previously (11, 12). The sequence of HI_0990 in each isolate was retrieved from these earlier sequence data. Variant calling and annotation were conducted using Bcftools and snpEff, respectively.

Polymerase chain reaction (PCR) and quantitative reverse transcription PCR (qRT-PCR)

H. influenzae isolates were incubated in brain–heart infusion broth supplemented with 15 mg/L of hemin and 15 mg/L of nicotinamide adenine dinucleotide, referred to as sBHI, to OD₆₀₀ = 0.6. The purification of the genomic DNA of each isolate has been described previously (11). Total RNA from the isolates was extracted using REzol C&T (Protech Technology, Taipei, Taiwan) and treated with TURBO DNase (Thermo Fisher Scientific, Hampshire, UK). Five micrograms of purified RNA were reverse transcribed using the Superscript III First-Strand Synthesis System (Thermo Fisher Scientific). PCR was performed using Power SYBR Green PCR Master Mix (Thermo Fisher Scientific) and StepOnePlus Real-Time PCR Systems (Thermo Fisher Scientific). The primers used are listed in Table S1. The primers used to detect igaB genes in H. influenzae isolates were as described in a previous study (9).

IgA1 protease activity assay

The H. influenzae isolates were cultured in sBHI to OD₆₀₀ = 0.6, and 100 µL of serial two-fold dilutions of culture supernatants were incubated with 10 µg/mL of native human IgA1 protein (Abcam, Cambridge, UK) at 37°C and gently shaken for 6 h. Culture supernatants alone and sBHI broth with human IgA1 protein were used as controls for the activity assay. The reactions were stopped using a protease inhibitor cocktail (Thermo Fisher Scientific). Polystyrene High-Bind Microplates (Corning, Corning, NY, USA) were coated with rabbit anti-human IgA-crystallisable fragment antibody (Exalpha Biologicals, MA, USA) at 4°C overnight, washed using 1 × phosphate-buffered saline with 0.05% Tween 20, and blocked using a protein-free blocking buffer (Thermo Fisher Scientific) at room temperature for 2 h. The protease reaction solutions were added to wells and incubated at 4°C overnight. The plates were then washed three times and incubated with horseradish peroxidase-conjugated mouse anti-human lambda light chain antibody (Abcam) at room temperature for 2 h. Tetramethyl benzidine and 2N sulfuric acid were used as the colour-developing substrate and cessation solution, respectively. Absorbance was measured at an optical density (OD) of 450 nm. Protease activity was calculated as the mean Δ OD 450 nm value (sample–culture medium) at various dilution folds. A change in IgA1 protease activity was defined as a 30% increase or decrease in the mean Δ OD 450 nm value compared with that of the Rd KW20 strain.

Structure-based protein stability analysis

The protein sequence and structure of IgA1 protease from the H. influenzae Rd KW20 strain (Uniprot accession #P44969; AlphaFold Database accession #AF44969-F1) were used as reference standards (13). The difference in folding free energy (ΔΔG or DDG) between the reference IgA1 protease and those with amino acid substitutions was employed to assess the structural stability of IgA1 protease. The ΔΔG values of IgA1 protease with different substitutions were calculated using the mCSM (14), DDMut (15), DynaMut2 (16), and ELASPIC2 (17).

Site-direct mutagenesis and transformation

Full-length iga from the Rd KW20 strain was cloned into the pMXLK_pFa vector (Addgene, Watertown, MA, USA) at the 5′ HindIII site and the 3′ BamHI site. The IgA1 protease open reading frame with amino acid substitutions was generated by site-directed mutagenesis, as described previously (11). The primers used are presented in Table S1. The methylated parental DNA template was digested using DpnI. Purification of plasmid DNA from DH5α cells and transformation of H. influenzae have been described previously (11).

Statistical analyses

SPSS 18.0 for Windows was utilized for statistical analyses. Chi-squared tests, receiver operator characteristic (ROC) curves, and logistic regression analyses were conducted to identify the association between amino acid substitutions and IgA1 protease activity. The Kruskal–Wallis test was used to compare continuous variables for three groups. Significance is set at P < 0.05 (two-tailed). Venn diagrams were generated using InteractiVenn (http://bioinfogp.cnb.csic.es/tools/venny/) (18). Phylogenetic analyses were conducted using Molecular Evolutionary Genetics Analysis version 11 (19). Evolutionary history was inferred using the maximum likelihood method with a bootstrap value of 100 and the Jones-Taylor-Thornton matrix-based model (20). A heatmap was generated using TBtools-II (21).

RESULTS

Expression profiles and activity of IgA1 protease in H. influenzae

According to comparisons with the IgA1 protease activity of the reference Rd KW20 strain, four isolates (A0007, A0650, A0765, and A3275) exhibited lower activity levels, 28 exhibited comparable activity levels, and 19 exhibited higher activity levels (Fig. 1). The IgA1 protease activity in the isolates was not correlated with the type or source of the specimen (Fig. S1A and B). The results of qRT-PCR revealed a similar expression level of iga (HI_0990) in all isolates and no igaB in any isolates (data not shown).

Fig 1.

IgA1 protease activities in Haemophilus influenzae isolates. Relative IgA1 protease activity of each isolate compared with the Rd KW20 strain is presented. A change in IgA1 protease activity was defined as a 30% increase or decrease in the mean value compared with that of the Rd KW20 strain. The mean levels of IgA1 protease activity in groups exhibiting a decrease (n = 4), a negligible change (n = 28), and an increase (n = 19) in different dilution folds of culture medium are indicated by red, gray, and blue lines, respectively.

Sequence analysis of H. influenzae iga

The sequence of iga in the H. influenzae isolates was retrieved from the whole-genome sequence data and compared with that from the Rd KW20 strain. The distribution of deduced amino acid substitutions in the IgA1 protease of each isolate revealed that the signal sequence (amino acids 1–25) was highly conserved, with only one substitution, T20A, detected in 14 isolates (Fig. 2A). The amino acids 61–70, 191–200, and 321–337 in the protease domain (amino acids 26–337), 351–360, 551–580, 801–810, 841–920, 1001–1190, and 1201–1409 in the nonactive site of the passenger domain (amino acids 338–1409), and 1521–1550 and 1571–1600 in the β-barrel domain (amino acids 1410–1694) were frequently substituted. The mean substitution rate of the protease domain among the isolates was 4.6%, with four isolates (A3052, A3423, A3574, and A3680) higher than 10% (Fig. 2B). The mean substitution rates of the first (amino acids 338–695), second (amino acids 696–1052), and final (amino acids 1053–1409) third of the nonprotease passenger domain were 5.2%, 8.2%, and 13.5%, respectively. The mean substitution rate of the β-barrel domain was less than 5%. Phylogenetic analyses revealed that changes in IgA1 protease were correlated moderately with the protease domain and strongly with the nonprotease passenger domain (Fig. 2B). Furthermore, the IgA1 protease activity was not correlated strongly with the phylogeny of the protein sequence, suggesting an involvement of multiple substitutions in determining the protease activity.

Fig 2.

Amino acid substitution rates of the IgA1 protease in the Haemophilus influenzae isolates. (A) Distribution of substitution rates per 10 amino acids of IgA1 protease in the Haemophilus influenzae isolates (n = 51) and (B) phylogenetic characteristics and heatmaps showing the amino acid substitution rates in different regions of H. influenzae IgA1 protease are presented. The sequence of IgA1 protease of the H. influenzae Rd KW20 strain was used as the reference.

Genetic variations and protein changes associated with IgA1 protease activity

Chi-squared tests revealed that 32 genetic variations, of which four were located in the protease domain, 26 in the nonprotease passenger domain, and two in the β-barrel domain, were associated with the IgA1 protease activity (Table S2). Two of these variations and their deduced amino acid substitutions, namely Asn914Lys and Lys1542Gln, were associated with an increase in the IgA1 protease activity, whereas the other 30 substitutions were associated with a decrease in the IgA1 protease activity. Asn352Lys-Pro353Ala, Leu912Phe-Ser913Thr, and Ser1187Asn-Asp1404Tyr cosubstitutions were observed. The Asn352Lys-Pro353Ala cosubstitution was strongly associated with Lys356Asn, whereas Gln916Lys, Gly917Ser, and Asn1008Asp exhibited significant associations. Logistic regression and ROC analyses revealed the association of Asn352Lys, Pro353Ala, Lys356Asn, Gln916Lys, Gly917Ser, Asn1008Asp, Lys1050Gln, and Asn1255Lys with a decreased IgA1 protease activity, whereas Asn914Lys with an enhanced IgA1 protease activity (Table 1). Overall, isolates containing multiple substitutions associated with decreased IgA1 protease activity possessed a low IgA1 protease activity. By contrast, 23 isolates containing the Asn914Lys substitution but not those associated with decreased IgA1 protease activity had high levels of IgA1 protease activity (Fig. 3).

TABLE 1.

Univariate logistic regression and ROC analyses of amino acid substitutions associated with IgA1 protease activity in H. influenzae isolates (n = 51) ^a

Variant	Logistic regression		ROC curve
	Odds ratio (95% CI)	P-value	AUROC (SE)	Sensitivity (%)	Specificity (%)	P-value
Protease domain
Glu142Asp	0.028 (0.000–2.295)	0.112	0.774 (0.104)	60	87	0.046
Tyr143Phe	0.026 (0.000–1.555)	0.08	0.778 (0.088)	100	46.7	0.028
Ala182Pro	0.000 (0.000–1.929)	0.06	0.926 (0.042)	100	87.2	0.005
Ser268Arg	0.129 (0.008–2.179)	0.155	0.689 (0.105)	100	37.8	0.136
Non-protease passenger domain
Asn352Lys	0.080 (0.009–0.711)	0.023	0.755 (0.077)	76.9	68.4	0.006
Pro353Ala	0.080 (0.009–0.711)	0.023	0.755 (0.077)	76.9	68.4	0.006
Lys356Asn	0.059 (0.006–0.566)	0.014	0.776 (0.072)	78.6	70.3	0.003
Ser555Asn	0.145 (0.007–2.921)	0.208	0.683 (0.124)	40	97.8	0.183
Lys613Thr	0.103 (0.005–2.049)	0.136	0.719 (0.102)	66.7	71.1	0.085
Thr910Ala	0.295 (0.055–1.580)	0.154	0.673 (0.096)	58.3	74.4	0.072
Leu912Phe	0.157 (0.021–1.178)	0.072	0.718 (0.093)	63.6	75	0.028
Ser913Thr	0.157 (0.021–1.178)	0.072	0.718 (0.093)	63.6	75	0.028
Asn914Lys	113.855 (3.541–366E03)	0.007	0.840 (0.059)	84.6	68.4	<0.001
Gln916Lys	0.017 (0.001–0.470)	0.016	0.816 (0.067)	72.7	77.5	0.001
Gly917Ser	0.028 (0.001–0.645)	0.026	0.790 (0.073)	90	56.1	0.005
Thr997Ile	0.000 (0.000–3.079)	0.084	0.862 (0.077)	100	63.8	0.017
Pro1006Thr	0.017 (0.000–2.197)	0.1	0.787 (0.106)	60	87	0.037
Asn1008Asp	0.078 (0.009–0.651)	0.018	0.753 (0.075)	78.6	67.6	0.006
Ser1049Asn	0.161 (0.022–1.191)	0.074	0.673 (0.087)	100	42.5	0.082
Lys1050Gln	0.089 (0.014–0.586)	0.012	0.727 (0.072)	100	44.1	0.009
Thr1067Arg	0.055 (0.002–1.774)	0.102	0.733 (0.096)	83.3	66.7	0.066
Ala1068Ser	0.055 (0.002–1.774)	0.102	0.733 (0.096)	83.3	66.7	0.066
Gly1071Glu	0.779 (0.234–2.585)	0.683	0.546 (0.085)	19	100	0.579
Glu1096Lys	0.055 (0.002–1.774)	0.102	0.733 (0.096)	83.3	66.7	0.066
Ala1109Asp	0.087 (0.001–7.025)	0.276	0.774 (0.187)	66.7	97.9	0.128
Ser1187Asn	0.000 (0.000–8.53E37)	0.374	0.980 (0.020)	100	98	0.103
Lys1189Arg	0.000 (0.000–3.95E07)	0.308	0.960 (0.028)	100	96	0.118
Asn1255Lys	0.084 (0.011–0.678)	0.02	0.745 (0.073)	85.7	59.5	0.007
Val1400Ile	0.025 (0.000–3.844)	0.151	0.803 (0.124)	75	83	0.046
Asp1404Tyr	0.000 (0.000–8.53E37)	0.374	0.980 (0.020)	100	98	0.103
β-barrel domain
Lys1542Gln	1.53E17 (0.000–6.99E38)	0.12	0.990 (0.014)	98	100	0.02
Thr1613Ile	0.000 (0.000–2.97E04)	0.12	0.990 (0.014)	100	98	0.02

^a ROC analyses show the power of each substitution for differentiating the IgA1 protease activity. Abbreviations: AUROC. Area under ROC curve; CI, confidence interval; ROC, receiver operating characteristic; SE, standard error.

Fig 3.

Associations of amino acid substitutions with IgA1 protease activity. A Venn diagram for Haemophilus influenzae isolates with different substitutions and IgA1 protease activities is presented. Data are presented as isolate name (relative IgA1 protease activity compared with that of the Rd KW20 strain).

Effects of substitutions on stability and activity of IgA1 protease

Four prediction tools, mCSM, DDMut, DynaMut2, and ELASPIC2, were used to evaluate the effects of 9 substitutions on the protein stability of IgA1 protease. The results revealed that Asn352Lys, Pro353Ala, Lys356Asn, Gln916Lys, and Gly917Ser substitutions tended to destabilize IgA1 protease, whereas Asn914Lys tended to stabilize this protein (Table 2). The evaluation results of these tools for Asn1008Asp, Lys1050Gln, and Asn1255Lys substitutions were inconsistent. Furthermore, functional complementation assays verified the effect of the Asn914Lys substitution on the enhancement of IgA1 protease activity in the Rd KW20 strain (Fig. 4).

TABLE 2.

Protein stability analyses of amino acid substitutions for IgA1 protease activity in Haemophilus influenzae isolates ^a

Variant	mCSM		DDMut		DynaMut2		ELASPIC2
	DDG	Interpretation	DDG	Interpretation	DDG	Interpretation	Score	Interpretation
Asn352Lys	−0.111	Destabilizing	−0.47	Destabilizing	−0.12	Destabilizing	−1.077	Destabilizing
Pro353Ala	−1.633	Destabilizing	−0.46	Destabilizing	−1.46	Destabilizing	−2.465	Destabilizing
Lys356Asn	−0.413	Destabilizing	−0.41	Destabilizing	−0.05	Destabilizing	−1.852	Destabilizing
Asn914Lys	0.146	Stabilizing	0.52	Stabilizing	0.19	Stabilizing	0.231	Stabilizing
Gln916Lys	−0.144	Destabilizing	−0.06	Destabilizing	−0.01	Destabilizing	−1.054	Destabilizing
Gly917Ser	−1.154	Destabilizing	−1.55	Destabilizing	−0.80	Destabilizing	−0.359	Destabilizing
Asn1008Asp	−0.144	Destabilizing	0.07	Stabilizing	−0.03	Destabilizing	−1.055	Destabilizing
Lys1050Gln	−0.131	Destabilizing	0.08	Stabilizing	0.02	Stabilizing	−0.800	Destabilizing
Asn1255Lys	−0.053	Destabilizing	0.06	Stabilizing	−0.11	Destabilizing	−1.054	Destabilizing

^a DDG, difference in folding free energy (kcal/mol). In ELASPIC2, the results are shown as scores, which reverse DDG values.

Fig 4.

Functional complementation assays. Relative IgA1 protease activity in the Haemophilus influenzae Rd KW20 strain after the expression of IgA1 protease with Asn352Lys-Pro353Ala-Lys356Asn, Gln916Lys-Gly917Ser, or Asn914Lys substitutions is presented. Data were obtained from three independent experiments. P values were obtained from student’s t tests. *P < 0.05; **P < 0.01.

DISCUSSION

IgA1 comprises approximately 90% of the upper respiratory IgA pool in humans and is crucial for antibacterial defense, such as protecting mucous membranes, inhibiting microbial adhesion to host epithelial cells, neutralizing microbial toxins and enzymes, and initiating potent effector functions of crystallisable fragment alpha receptor I-bearing macrophages, thereby eliminating invading pathogens (22 – 24). H. influenzae IgA1 protease is a virulence factor that cleaves IgA1 proteins and promotes bacterial invasion and colonisation in hosts. In addition to the IgA1 protease, three autotransporters, namely Hap (Haemophilus adherence and penetration), Hia (H. influenzae adhesin.), and Hsf (Haemophilus surface fibrils), were identified in clinical H. influenzae strains. Hap contains a protease domain but does not cleave human IgA1 (25), whereas Hia and Hsf lack a protease domain. Hia, Hap, and Hsf facilitate the initial contact and augment the adherence of bacterial cells to the epithelium and extracellular matrices (26 – 33), whereas IgA1 protease shields bacteria from innate immune responses and promotes colonization and persistence in hosts. H. influenzae is the most common bacterial cause of the exacerbation of COPD (34, 35), and the gene pattern and expression of H. influenzae IgA1 protease are highly associated with this event (36). H. influenzae strains containing igaB exhibit a higher level of IgA1 protease activity and a more severe disease stage of COPD than those that do not, which might result from enhanced intracellular survival in epithelial cells and prolonged duration of carriage in the hosts (9, 10, 37). In addition to igaB, the variable expression pattern of H. influenzae IgA1 proteases, conferred by changes in iga gene sequences, influences the propensity of strains in the setting of COPD (36, 38). In this study, we did not detect the presence of igaB in our isolates and focused on addressing the association of iga polymorphism with the phenotype of IgA1 protease activity in H. influenzae.

A crystallographic study by Johnson et al. revealed that the passenger domain of H. influenzae IgA1 protease contains an N-terminal trypsin/chymotrypsin-like protease domain that includes a catalytic site with the sequence GDSGSPLF, and a β-helical spine as the major structural components (39, 40). Autoproteolytic cleavage of H. influenzae IgA1 protease occurs at a proline-rich sequence near the junction of the passenger domain and the β-barrel domain, resulting in the extracellular release of the passenger domain from the β-barrel domain that is embedded in the outer membrane (41 – 43). Meningococcal outer membrane protein 85 and Bam complex (BamA-E) in Yersinia enterocolitica are essential for the insertion and folding of β-barrel outer membrane proteins (44, 45). Although Bam homologs have been characterised in H. influenzae, their influence on IgA1 protease is poorly understood. Further studies are required to evaluate the effects of the H. influenzae Bam complex on IgA1 protease membrane localisation and the translocation of the passenger domain across the outer membrane.

Our analyses revealed that the amino acid substitution rates in the signal sequence, protease domain, and β-barrel domain of IgA1 protease were lower than 5%, suggesting that these domains are crucial for the protein functions. The residues of the nonprotease passenger domain exhibited greater variability, particularly residues 1053–1409. In total, 625 genetic variations with deduced amino acid substitutions were identified among the H. influenzae isolates. No substitutions occurred in the catalytic site (residues 286–293), and no significant changes in IgA1 protease activity were observed in these isolates. Two isolates, namely A1231 and A3289, contained the Ser1018Asn substitution, which is located in one of the autocleavage sites. However, the influence of the Ser1018Asn substitution on IgA1 protease activity was inconclusive because the IgA1 protease activity in Isolate A3289 was similar to that in the Rd KW20 strain, but was doubled in Isolate A1231. Statistical analyses revealed that eight substitutions were associated with a decrease and one with an increase in IgA1 protease activity. In H. influenzae IgA1 protease, Asn352Lys, Pro353Ala, and Lys356Asn were located close to the protease domain (residues 26–337), Asn914Lys, Gln916Lys, and Gly917Ser to one of the substrate binding domains (residues 564–657, 710–743, and 786–819), and Asn1008Asp and Lys1050Gln to the autocleavage sites of the IgA1 protease (39).

Amino acid substitutions can alter the folding free energy and the structural stability of proteins. We used various tools to calculate the DDG values of H. influenzae IgA1 protease with specific substitutions and assessed their effects on protein stability. mCSM utilizes machine learning to predict the effects of missense mutations through structural signatures derived from graph-based concepts (14). DDMut applies a Siamese network that integrates graph-based representations to capture atom distance patterns and calculate ΔΔG from point mutations (15). DynaMut2 combines normal mode analysis for capturing protein motion with graph-based signatures to assess the impact of amino acid substitutions on protein stability and dynamics (16). ELASPIC2 uses homology and machine learning models to predict the effects of mutations on protein folding and protein-protein interactions (17). These tools have been employed to predict the impacts of missense mutations on the stability of bacterial, viral, or human proteins (46 – 51). In our analyses, Asn352Lys, Pro353Ala, Lys356Asn, Gln916Lys, and Gly917Ser substitutions were predicted to destabilize IgA1 protease, whereas Asn914Lys was predicted to increase IgA1 protease stability. Although Asn352Lys, Pro353Ala, and Lys356Asn substitutions are not located in the catalytic site and the substrate recognition and binding subdomains (domain2: residues 564–657; domain 3: residues 710–743; domain 4: residues 786–819) (39), the destabilizing nature of these substitutions may diminish the capacity of H. influenzae IgA1 protease to cleave IgA1. The Asn914Lys substitution lies near Gln916Lys and Gly917Ser and possesses opposite effects to Gln916Lys and Gly917Ser on protein stability and protease activity. These substitutions in the nonactive site of the passenger domain may alter the β-helical spine structure of IgA1 protease, thereby indirectly affecting the binding affinity to IgA protein substrates or the proteolytic activity. Further crystallographic studies of IgA1 proteins complexed with H. influenzae IgA1 protease variants are required to better understand how these substitutions affect the phenotype of this protein.

In conclusion, this study reports the relevance of amino acid substitutions in the nonactive site of the passenger domain in the activity of H. influenzae IgA1 protease. Although the identified substitutions did not cause huge impacts on the catalytic or secretion activities of H. influenzae IgA1 protease in vitro, they may substantially affect bacterial infection microenvironments. Further studies are needed to evaluate the effects of these amino acid substitutions in IgA1 protease on the progression of H. influenzae diseases or the exacerbation of COPD. These insights enhance our collective understanding of the infection and immunomodulation in IgA1 protease-producing microbes.

ETHICS APPROVAL

This study was approved by the Institutional Review Board (No. 2023025) in accordance with the ethical standards noted in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The Review Board waived the requirement for obtaining informed consent because this retrospective study used only the bacterial samples and did not have any negative impact on the patients.

DATA AVAILABILITY

The data generated from this study can be available from the corresponding author on reasonable request. Sequencing reads of each H. influenzae isolate are available from the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA1057504).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/iai.00193-24.

Figure S1. iai.00193-24-s0001.tif.

Percentages of Haemophilus influenzae isolates (n = 51) with different IgA1 protease activity levels obtained from different (A) specimen types and (B) specimen sources.

Supplemental legend. iai.00193-24-s0002.docx.

Legend for Fig. S1.

Table S1. iai.00193-24-s0003.docx.

Sequences of primers.

Table S2. iai.00193-24-s0004.docx.

Genetic variations and deduced amino acid substitutions associated with IgA1 protease activity in H. influenzae isolates (n = 51).

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.