Klebsiella pneumoniae, a Gram-negative bacterium of Enterobacterale, is ubiquitously distributed in natural environment. It is typically armed with a polysaccharide capsule along with type 1 and 3 fimbriae, but lacks flagella. As an opportunistic pathogen, K. pneumoniae is associated with a range of infections including endocarditis, and sepsis (Navon-Venezia et al., 2017). Hypervirulent K. pneumoniae (HvKP) strains can even cause fatal invasive syndromes such as necrotizing, lung abscess, pyogenic liver abscess, and endogenous endophthalmitis (Paczosa et al., 2016).

The first carbapenem-resistant K. pneumoniae (CRKP) strain was reported in the United States in 1996 (Yigit et al., 2008), which has become increasingly prevalent worldwide. At present, the proportion of carbapenem resistance in clinical K. pneumoniae strains is around 25.0% in China mainland (Li et al., 2023). CRKP commonly exhibits multidrug-resistance, including resistance to quinolones, aminoglycosides, and cephalosporins. These resistance profiles confer prolonged hospital stay, increased healthcare cost, and elevated mortality (Xu et al., 2017), which prompts the World Health Organization to classify CRKP as a priority pathogen.

The convergence of antimicrobial abuse and the continual evolution of K. pneumoniae have facilitated the emergence of superbugs with carbapenem resistance and hypervirulence (Martin and Bachman, 2018). In China mainland between 2015-2017, over 5.0% of carbapenem-resistant isolates demonstrated the dual risks, with a rising trend (Zhang et al., 2020). This ‘superbug’ poses a formidable challenge for clinicians. Drug resistance may be developed by K. pneumoniae through various mechanisms, such as hydrolytic enzymes, loss of outer membrane proteins (OMPs), over-expression of efflux pumps, topoisomerases, lipopolysaccharide modifications and so on (Hennequin and Robin, 2016; Aguilar et al., 2025; Uehara et al., 2025b). Among these, the absence or reduced expression of the porins OmpK35 and OmpK36 significantly restricts the penetration of carbapenems into the cell, thereby conferring resistance (Zhang et al., 2015; Park et al., 2025; Xie et al., 2025). There are also studies indicating that mutations, defects, or loss of function in OmpK35/36, if further combined with KPC or CTX-M, may even lead to an increase in the minimum inhibitory concentration (MIC) of ceftazidime-avibactam (CZA) (Nelson et al., 2017; Shen et al., 2017; Castanheira et al., 2020; Gaibani et al., 2020; Zhao et al., 2024).

OmpK35 and OmpK36 are major trimeric β-barrel porins in the outer membrane of K. pneumoniae and facilitate the passive diffusion of small hydrophilic molecules, including nutrients and certain antibiotics like β-lactams (Brunson et al., 2019). They share structural and functional similarities with OmpF and OmpC in Escherichia coli respectively, and play critical roles in maintaining membrane integrity alongside peptidoglycan and lipopolysaccharide layers (Sugawara et al., 2016; Davin-Regli et al., 2024). However, despite their importance, a systematic review of OmpK35/36 impact on K. pneumoniae lacks.

To address this gap, an epidemiological analysis of OmpK35/36 sequences available in the GenBank database was conducted in this paper. Furthermore, the serotype K1 strain NTUH-K2044 was used to elucidate the impacts of OmpK35/36 on its drug-resistance, particularly resistance to carbapenems and CZA.

As confirmed in the study, OmpK35 and OmpK36 and their corresponding genes exhibited distinct characteristics in K. pneumoniae. OmpK35/36 loss is a contributor but not determinant in the formation of carbapenem-resistance without bla_KPC-2, as well as in the formation of CZA-resistance with one bla_KPC-2.

Gene ompK35/36 was widely present in K. pneumoniae strains (Figure 1A). Given the high sequence similarity of OmpK35 and OmpK36 (Figure 2) in K. pneumoniae strains, they may become potential targets for antibacterial drugs and vaccine development. A notable proportion of K. pneumoniae strains with defect in ompK35/36 were found to carry carbapenemase genes (Figure 1D). Since this analysis was conducted at the genomic level without assessing transcription or translation, the actual prevalence of such defects may be substantially underestimated. Previous studies have reported that CRKP frequently harbors point mutations, insertions, or truncations in ompK35/36, leading to reduced or absent porin expression, which in turn enhances resistance to β-lactam antibiotics (Bongiorno et al., 2023; Hamzaoui et al., 2018; Joshi et al., 2021; O'Neill et al., 2025). The deficiency of OmpK35/36 in K. pneumoniae, combined with the production of diverse β-lactamases (including carbapenemases), has posed a great challenge to the therapy based on carbapenems. Although the genomic data from GenBank may be influenced by regional research biases and may not fully reflect actual clinical distribution, a deep survey of ompK35/36 prevalence can provide a foundational understanding, which is essential for subsequent functional investigation. It is noted that antibiotics and phages drive region-specific diversity of OmpK36 in K. pneumoniae and such diversity confers some influence on drug-resistance of K.pneumoniae (Cai et al., 2018; Dunstan and Bamert, 2023; Royer et al., 2025a). However, structure variations of OmpK36 were not included in this study.

Functional studies were conducted using the reference strain NTUH-K2044 to elucidate the role of OmpK35/36 in antimicrobial resistance. In this strain, expression level of porin genes varied considerably: ompK36 was predominant, and ompK35 was moderately expressed, while ompK26 and ompK37 remained largely silent. Under hyperosmotic stress (2.5% NaCl), the expression balance of ompK35/36 in K. pneumoniae shifted, indicating that the expression of porins was regulated by osmotic pressure. It was consistent with the research results of porins in Escherichia coli: the two-component system of EnvZ/OmpR could participate in the regulation of bacterial response to external osmotic pressure stimuli, by regulating the expression of OmpF and OmpC (Kenney and Anand, 2020). In the high-osmolarity intestinal environment, ompK36 expression was favored. Since OmpK35 formed a larger pore than OmpK36 (Sugawara et al., 2016), its downregulation would help bacteria maintain osmotic homeostasis. Clinically, many isolates exhibited a shift from OmpK35 to the smaller-pored OmpK36, reducing antibiotic uptake and enhancing bacterial survival (Bongiorno et al., 2023; Rocker et al., 2020).

Antibiotic susceptibility tests revealed that OmpK35/36 loss primarily affected resistance to β-lactams (Table 2). Deficiency in one or both porins increased MICs of various β-lactams (Srinivasan et al., 2012; Asare Yeboah et al., 2025; Birgani et al., 2025). The impact was found more in ΔompK36 than in ΔompK35, which was consistent with the dominant expression of OmpK36 (Yasmin et al., 2024) in NTUH-K2044. But if ompK35 was further deleted, it would result in even higher drug-resistance compared to single mutants, meaning that ΔompK35/36 exhibited higher antibiotic resistance than ΔompK35 or ΔompK36. For ΔompK35/36, the MICs of cefoxitin, cefazolin, and meropenem reached 8, 8, and 4 times compared to those of ΔompK36 respectively. Complementation only partially restored susceptibility, which was likely due to limited OmpK36 re-expression (20 ~ 30% of wild-type levels, see Supplementary). The different effect of OmpK35/36 on various β-lactams may stem from difference in antibiotic size, structure and charged properties, which would influence passage through the porin channels (Matovina et al., 2021; Acharya et al., 2024; Royer et al., 2025b). Therefore, significant difference in molecular weights, three-dimensional structures and charged properties among various β-lactams may be the core reasons for the varying degrees of impacts of OmpK35/36 deficiencies on drug-resistance.

Moreover, the loss of OmpK35/36 had a very limited impact on resistance to carbapenems and did not result in the formation of CRKP (Table 2). The antimicrobial susceptibility tests of conjugants are shown in Table 3. Firstly, it should be emphasized that the interaction between HvKP and drug-resistance plasmids after conjugation is not discussed here. Some literature suggests that HvKP strains carrying bla_KPC-positive plasmids (K1/K2 HvKP strains acquire carbapenem-resistance plasmids) were difficult to simultaneously possess hypervirulence and carbapenem-resistance, and both may decrease when the two were combined (Tian et al., 2022). In this paper, the restoring sensitivity of conjugant NTUH-K2044+bla_KPC-2 to aztreonam, cefepime, and so on, seemed to verify this. However, this study was conducted under the same basic conditions for antimicrobial susceptibility tests of the conjugants, with the only difference being whether OmpK35/36 exists. Then, it can be concluded from Table 3 that the deletion of OmpK35/36, particularly the double deletion, greatly helped NTUH-K2044+bla_KPC-2 induce resistance to β-lactams. It can be also observed that the absence of OmpK35/36 can increase the MIC of CZA against NTUH-K2044+bla_KPC-2; single OmpK35/36 loss doubled the MIC and double loss quadrupled the MIC, which is different from another study (Nelson et al., 2017). Nevertheless, the MIC of CZA against ΔompK35/36+bla_KPC-2 (1/4 mg/L) was far away from resistance breakpoint (16/4 mg/L). Recent study (Uehara et al., 2025a) investigated the impact of porin deletions on cefepime-taniborbactam activity against K. pneumoniae and found taniborbactam didn’t have a strong dependence on OmpK35 or OmpK36 for periplasmic accumulation. In all, the deficiency of porin proteins has a promoting effect on other resistance mechanisms of K. pneumoniae but it is not a determinant of carbapenem or CZA resistance.

In summary, the conservation of the ompK35/36 and OmpK35/36 provided a possible molecular basis for the development of new targeted antibacterial drugs and vaccines, which was beneficial for better prevention and treatment of K. pneumoniae infections. The absence of OmpK35/36 could lead to increased resistance of strains to certain β-lactams. Although OmpK36 was the dominant porin, OmpK35 could play a synergistic role in the drug-resistance of K. pneumoniae while they were both absent. The absence of OmpK35/36 alone or both cannot determine the formation of carbapenem or CZA resistance, which only plays an auxiliary role in such formation.