Use of trans-complementation method to determine the effects of various ftsI mutations on β-lactamase-negative ampicillin-resistant (BLNAR) Haemophilus influenzae strains

Haemophilus influenzae is a causative agent of serious infections, especially among children. β-lactam antibiotics are commonly used for the treatment of these infections. Among H. influenzae isolates, β-lactam resistance is due to the presence of β-lactamase, or to mutations in the ftsI gene that generate altered PBP3 (penicillin-binding protein 3) with reduced affinity for β-lactams (BLNAR—β-lactamase-negative, ampicillin-resistant). Wild-type ftsI gene encoding for PBP3 was amplified in whole from β-lactam susceptible H. influenzae Rd and cloned in pLS88 plasmid to obtain pADUTAS17, which was then used to transform known BLNAR strains, susceptible strains, and a strain (CF55) with wild-type ftsI but unexplained reduced β-lactam susceptibility. Ampicillin and cefotaxime MICs (minimum inhibitory concentration) were determined after transformation with pLS88 and pADUTAS17 plasmids. The results showed that antibiotic susceptibilities were not affected by trans-complementation for isolates carrying wild-type ftsI gene. However, trans-complementation for all BLNAR strains showed decreases between − 0.957 and 0.5-fold for ampicillin and cefotaxime, confirming the role of the PBP3 substitutions in the BLNAR phenotype of these isolates. The first article showed that trans-complementation might be a useful tool in the investigation of decreased β-lactam susceptibility in H. influenzae.


Introduction
Haemophilus influenzae is part of the normal flora of the respiratory tract and an important opportunistic pathogen. Some strains have a polysaccharide capsule (serotypes a-f) and can cause invasive diseases such as meningitis and septicemia, especially among infants and children under 5 years old (Oliver et al. 2021). There are two broad groups of H. influenzae. Some strains have a polysaccharide capsule (serotype a-f) but the majority of strains seen in non-invasive disease lack the capsule locus, and are therefore termed as non-typable H. influenzae (NTHi) but they are still identified as NTHi based on standard phenotypic and genotypic criteria. These strains tend to cause non-invasive opportunistic infections such as otitis media and communityacquired pneumonia (Tristram et al. 2007).
The first choice for the treatment of H. influenzae is β-lactam drugs. However, β-lactamase-negative ampicillinresistant H. influenzae (BLNAR) and β-lactamase-producing ampicillin-resistant H. influenzae (BLPAR) are a problem for the treatment of H. influenzae with β-lactam drugs. BLNAR shows penicillin resistance and BLPAR is resistant to cephalosporins and penicillin. BLNAR strains are common in all countries, especially in Japan and Korea (Yamada et al. 2020).
The role of specific amino acid substitutions in penicillin-binding protein 3 (PBP3) in producing a BLNAR phenotype has been thus far demonstrated by ftsI sequencing and deduced PBP3 amino acid substitutions of large numbers of strains with a BLNAR phenotype, by transforming susceptible strains with ftsI genes from BLNAR Communicated by Erko Stackebrandt. strains, and by utilizing site-directed mutagenesis of the ftsI gene to generate desired substitutions. β-Lactam resistance in H. influenzae can be caused by the production of the plasmid-mediated TEM-1 β-lactamase enzyme which is the most common type of β-lactamase, or less common type ROB-1 β-lactamase, and in both instances, resistance is limited to the aminopenicillins such as ampicillin. Alternatively, in the BLNAR strain, the resistance is due to mutations in the ftsI gene that generate altered PBP3 with reduced affinity for all β-lactams. Consequently, these isolates have raised amoxicillin/clavulanate and cephalosporin MICs (minimum inhibitory concentration) in addition to high ampicillin MICs (Hacıseyitoğlu and Sümerkan 2021).
BLNAR strains were initially categorized into three groups by Ubukata et al (2001) based on specific amino acid substitutions in the transpeptidase region of PBP3. The most important substitutions are the BLNAR defining R517H (Group I) or N526K (Group II), whereas some other substitutions enhance BLNAR resistance, such that group III strains typically have M377I, S385T, and L389F in addition to N526K and higher β-lactam MICs. Many other PBP3 substitutions have now been identified but their contributions to resistance are unclear (Tristram et al. 2007;Ubukata et al. 2001).
While β-lactamase production and altered PBP3 are the most common and well-characterized β-lactam resistance mechanisms in H. influenzae, other mechanisms such as mutations to the AcrAB efflux pump or changes to outer membrane permeability have been reported in rare instances (Kaczmarek et al. 2004). While demonstration of altered PBP3 by sequencing of the ftsI gene in a resistant isolate might confirm the presence of a known resistance mechanism, it does not exclude the presence of an additional mechanism. However, if observed resistance is not reversed by trans-complementation with normal ftsI to a range seen in wild-type strains, then another contributing mechanism is suggested. Thus, trans-complementation with normal ftsI might be a useful tool for the investigation of β-lactam resistance in H. influenzae.
Amplified ftsI gene was ligated with pLS88 restricted with PstI using T4 DNA ligase enzyme (Thermo Scientific, ABD) overnight at 16 °C. The ligation mixtures were transformed into competent DH10B E. coli to generate pADUTAS17 (Sambrook et al. 2001).

Transformation of BLNAR strains with pADUTAS17
Various strains of H. influenzae (Table 1) were then transformed with either pLS88 or pADUTAS17. These include three strains of Rd previously transformed with mutated ftsI (RdΏBLNAR1, RdΏBLPACR4, and RdΏBLPACR7), a control strain (ATCC 49247), and a clinical isolate (UTAS 252)-all with BLNAR phenotype and genotype, a control strain (Rd, ATCC 51907) with BLNAS phenotype and genotype, and a clinical isolate (CF55) with a BLNAR phenotype but BLNAS genotype (absence of BLNARassociated mutations in ftsI). Competent cell preparation and transformation of BLNAR strains were performed using electroporation as previously described by Mitchell et al. (1991), using cooled 2 mm electroporation cuvettes, and the electroporator (BioRad micropulser) was set at 2.5 kV. Selection of positive colonies was made using BHI supplemented with vitox, 15 µg/mL NAD, and 15 µg/mL haemin concentrations containing 50 µg/mL kanamycin (Thermo Scientific, ABD) (Ubukata et al. 2001).

Determination of MIC concentration of wild-type and transformant BLNAR strains against the β-lactam antibiotics
MIC concentrations of both wild-type and transformant BLNAR strains (Table 1) containing plasmid-expressed non-mutated ftsI gene were determined to analyze the change in MIC concentrations against β-lactam antibiotics, cefotaxime, and ampicillin. For this purpose, E-Test (OXOID, UK) method was used according to the manufacturer's instructions.

Results
In this study, a total of seven H. influenzae strains were used to transform with pLS88 plasmid and pADUTAS17 plasmid which contained non-mutated ftsI gene in pLS88 plasmid, and the changes in susceptibility to ampicillin and cefotaxime antibiotics were analyzed with and without transcomplementation with non-mutated ftsI gene. These strains were H. influenzae RdΏBLNAR1 (Matic et al. 2003), H. influenzae RdΏBLPACR4 (Matic et al. 2003), H. influenzae RdΏBLPACR7 (Matic et al. 2003), H. influenzae CF55 (Tristram et al. 2007), H. influenzae UTAS252 (Søndergaard et al. 2015), H. influenzae ATCC49247, and H. influenzae Rd strains (ATCC51907) ( Table 1). We chose cloned strains (Rd transformed with mutated ftsI) because we have a baseline MIC (that of plain Rd) to see if the MIC of the cloned Rd BLNAR is close to totally reversed by trans-complementation. Then clinical or control strains with a variety of PBP3 substitutions were selected to demonstrate the utility of the model. Also, a strain (CF55 strain) was used that had a high ampicillin MIC but wild-type ftsI as a control in our model. In addition, we used plain pLS88 to transform all our strains to show that any reversion of resistance is due to the ftsI and not anything else on pLS88. H. influenzae Rd (ATCC51907) and CF55 strains have no mutation and the remaining five strains had mutations in the ftsI gene. The MIC results for cefotaxime and ampicillin antibiotics of the transformants are shown in Table 2. Transformants with pLS88 are shown as − 1 and with pADUTAS17 as − 2. Used H. influenzae names, resistance phenotypes, and found mutations in strains were explained. Resistance phenotype refers to the wild-type strains. '−1' means to transform of pLS88 plasmid and '−2' refers to transform of pADUTAS17 which contains wild-type ftsI in pLS88 plasmid. In table, some strains have mutations in the ftsI gene, and these mutations have been described. BLNAS (β-lactamase-negative ampicillinsensitive), BLNAR (β-lactamase-negative ampicillin-resistant) and BLPACR (β-lactamase-positive amoxicillin/clavulanate-resistant) According to the results, the MIC value of Rd transformant with wild-type ftsI gene for ampicillin did not show any difference when compared to Rd strain transformant with pADUTAS17, and the MIC value for cefotaxime showed a 0.5-fold decrease between transformed Rd strain transformed pLS88 plasmid and Rd strain with ftsI gene. H. influenzae CF55-2 did not show a significant change in sensitivity for both antibiotics. UTAS252 strain showed − 0.812-and − 0.789-fold decrease in ampicillin and cefotaxime MICs, respectively, after transformation with pADUTAS17 plasmid. RdΏBLNAR1-2 transformant strain showed increased sensitivity to − 0.81-fold decrease MIC value for ampicillin antibiotic and − 0.872-fold decrease MIC value for cefotaxime antibiotic. RdΏBLPACR4-2 strain enhanced sensitivity as − 0.81 times for ampicillin and -0.936 times for cefotaxime. RdΏBLPACR7-2 transformant showed a − 0.75-fold decrease in sensitivity for cefotaxime antibiotic. H. influenzae ATCC49247-2 showed − 0.937-and − 0.957-fold decrease for ampicillin and cefotaxime, respectively.

Discussion
The levels of β-lactam resistance or decreased susceptibility in H. influenzae due to altered PBP3 vary considerably depending on which amino acid substitutions are present. However, differences in β-lactam MICs are also often observed between isolates with identical substitutions and although this may be due to the imprecision of MIC testing or inherent baseline variation in susceptibility, it may also be due to the presence of additional unrecognized resistance mechanisms. Current investigative techniques such as ftsI sequencing and transformation studies may not be sufficient to detect the presence of additional unknown β-lactam resistance mechanisms. Our hypothesis is that the degree to which trans-complementation with wild-type ftsI restores susceptibility in a resistant strain might be useful to confirm that mutated ftsI is responsible for the observed resistance or alternatively, only contributing to the resistance and may indicate the presence of another mechanism. If, after trans-complementation with non-mutated ftsI, the decreased susceptibility of a strain can be reversed to the range seen in wild-type strains normal PBP3, then this suggests that the original decreased susceptibility was due to the altered PBP3. However, if the decreased susceptibility of a strain is either not reversed or only partially reversed following transcomplementation, then the presence of another mechanism is suggested and worthy of investigation.
The three RdΏ strains used in the present study were from Matic et al. (2003) and obtained after transformation of the Rd strain with mutated ftsI from three different BLNAR isolates. Transformation of these strains with non-mutated ftsI normally should revert β-lactam susceptibilities of these strains to Rd strain with pLS88 plasmid. After transformation, significant but not total reduction of the MIC of the Rd strain was seen for all strains. Incomplete reversion is not unexpected, because although non-mutated ftsI is introduced during trans-complementation and will encode normal PBP3, the mutated ftsI is still present and still encoding altered PBP3. The result will be a hybrid strain with both normal and abnormal PBP3 and some residual reduction in susceptibility. Strains ATCC 49247 and UTAS 252 both showed significant reduction of ampicillin and cefotaxime into the susceptible range, consistent with these strains having mutated ftsI-mediated resistance. CF55 strain is a clinical strain with a BLNAR phenotype, but without known BLNAR-associated ftsI mutations, and it did not show reversion of MIC value, so it is an example of the use of this approach to exclude altered PBP3 as a resistance mechanism in strains with uncharacterized reduced β-lactam susceptibility.
The significance of this study is that it shows that transcomplementation can be used to not only confirm the role of BLNAR-defining PBP3 substitutions in strains with decreased β-lactam susceptibility but also to suggest the presence of other mechanisms.

Conclusion
Our results showed increased susceptibilities for all strains to ampicillin and cefotaxime when trans-complemented with normal ftsI except H. influenzae Rd and CF55 strains because these strains had no mutation in the ftsI gene. In the investigation of decreased β-lactam susceptibility in H. influenzae trans-complementation might be a useful tool. This is the first article that showed resistance reversal using trans-complementation in BLNAR strains for the investigation of β-lactam resistance in H. influenzae. A future perspective with the results of the present study may be the determination of the effects of each mutation in PBP3 for susceptibilities to β-lactams other than the conserved sites.
Author contributions BB was responsible for the organization and coordination of the study. MY and ST were the investigators and were responsible for the data analysis. All authors contributed to the writing of the final manuscript.
Funding This study is supported by Adnan Menderes University BAP TPF11016. Melis YALÇIN is supported by 100/2000 YOK scholarship in Turkey.

Data availability
The ftsI gene sequence of H. influenzae Rd (ATCC51907) has been deposited in GenBank under the accessionnumber NC000907.