Structural Analysis and Protein Binding of Cephalosporins

Abstract

Cephalosporins are a widely used subclass of β-lactam antibiotics that demonstrate variable protein binding independent of generation or antibiotic coverage. Prior work analyzed carbon 3 (C3) and carbon 7 (C7) substituents (locations of R₂ and R₁ groups respectively) for protein binding interactions. This study builds upon these results with statistical analysis of additional agents of the class. Chemical structures of 23 cephalosporins were used to identify the presence of 40 functional groups, and correlative relationships were identified using established protein binding data. Four functional groups were significantly correlated with protein binding: tetrazole (positive association), pyridinium, primary amine, and quaternary amine (negative associations). Cephalosporins with a negative charge at physiological pH were associated with increased protein binding. Analysis of tetrazole-containing cephalosporins and ceftriaxone indicates the need for further study of the potential role in protein binding of neutral or negatively charged aromatic nitrogen heterocycles linked at the C3 position by a thiomethylene group.

As a determinant of distribution, protein binding of drugs is a critical pharmacokinetics consideration with therapeutic implications. Human serum albumin is the most abundant blood protein and the major determinant of drug–protein binding.¹ Many acidic and neutral drugs bind to serum albumin, a basic protein, through intermolecular interactions. While binding is reversible, the extent of binding determines the free fraction of drug that is available to distribute to the target site of action. Serum albumin concentrations are altered in several physiological states such as pregnancy, liver disease, and malnutrition, which changes the free fraction of drug for agents that are highly bound.¹ Therefore, in such situations, drug–protein binding interactions are relevant to drug therapy for both selection of an agent and assessment of patient response.

Protein binding data is readily available for cephalosporins, a subclass of the β-lactam class of antibacterial drugs. However, specific functional group interactions associated with protein binding patterns are not well characterized.² While cephalosporins are grouped into generations based on their antimicrobial spectrum of coverage, there is not a correlation between generation and extent of protein binding. This is illustrated by two highly protein bound cephalosporins that differ in generation and bacterial coverage, cefazolin and ceftriaxone.^3,4 Previous studies have analyzed structural features of cephalosporins and other drugs known to bind to albumin. For example, ceftriaxone was determined to bind to the same site on albumin as endogenous bilirubin, which may be associated with displacement of the drug.⁵ The binding interactions of cephalosporins with albumin are hypothesized to be ionic in nature, and intra-class variation is thought to be primarily associated with the R2 group on the C3 position of the cephalosporin pharmacophore, as shown in Figure 1.^6,7

Figure 1.

Cephem pharmacophore and ceftriaxone (R₁ purple, R₂ blue).

Cephalosporins may share identical or occupy proximal binding sites on albumin with warfarin, which binds in the subdomain IIA of site I, also referred to as Sudlow’s site I. Subdomain IIA of site I, shown in Figure 2, is a hydrophobic pocket with a hydrophilic, positively charged entrance; an ideal molecule or residue to bind to this pocket, such as warfarin, is small, lipophilic, and negatively charged.⁸

Figure 2.

Warfarin structure and bound to human serum albumin (negatively charged at physiological pH 7.4).

The cephem pharmacophore features a carboxylic acid group, which exists predominantly as a negatively charged carboxylate at physiological pH of 7.4. The auxophore moieties at the R₁ and R₂ positions feature ionizable functional groups for some members of the class. All cephalosporins exist as either negatively charged (mono- or dianions) or neutral zwitterions at physiological pH. Therefore, the wide variation of protein binding demonstrated by cephalosporins warrants investigation of the role of auxophore functional groups.

Clinical databases, designed to provide access to information for healthcare providers making clinical decisions, are also useful as repositories of large quantities of compiled clinical data. In addition to its accessibility, this data is useful in that it is derived from pooling of multiple sources and studies.⁹ Increased sampling augments the validity of its generalizable use through minimization of random error. Statistical analysis of clinical data and structural features of drugs affords the possibility of identifying correlations of potential relevance to scientists and healthcare providers.¹⁰ This study was therefore designed to compile clinical and chemical information to identify associations between degree of binding and presence of functional groups to provide additional information for future research on cephalosporin protein binding.

Methods

A cephalosporin was defined as any agent that binds to and inhibits transpeptidase and contains the cephem functional group. To be included in this analysis, drugs had pharmacokinetics data available through a drug information database or package insert, had a cephalosporin generic name, and were classified as a cephalosporin antibiotic according to the U.S. Food and Drug Administration (FDA) Established Drug Class (EDC) text phrases.¹¹ Drug candidates in clinical trials, molecules without generic drug names, or drugs lacking the necessary pharmacokinetics data were excluded. Cephalosporins from each generation were included. Generic names, extents of protein binding, and predicted in vivo charge states of cephalosporins analyzed in this study are contained in Table 1. Structures of cephalosporins are available in the Supporting Information.

Table 1. Cephalosporin Binding and Chemical Parameters.

cephalosporin	protein binding (%)	predicted net charge at pH 7.414	pKa (strongest acid)14	polar surface area (Å2)14
ceftriaxone	95	–2	2.7	208.98
cefoperazone	93	–1	3.19	220.26
cefotetan	88	–2	3.03	219.93
cefditoren	88	–1	2.27	160.10
cefazolin	80	–1	2.84	156.09
cefoxitin	79	–1	3.39	148.26
cefdinir	70	–1	2.73	158.21
cefixime	65	–2	2.54	184.51
ceftibuten	65	–2	2.85	38.49
cefuroxime	50	–1	2.96	173.76
cefotaxime	50	–1	2.73	173.51
cefprozil	36	0	2.64	158.26
cefpodoxime	33	–1	2.75	156.44
ceftizoxime	30	–1	2.66	147.21
cefaclor	25	0	2.83	112.73
cefepime	20	0	2.82	150.04
ceftaroline	20	–1	1.8a	223.24
cefadroxil	20	0	3.25	132.96
ceftolozane	20	0	2.49	302.21
ceftobiprole	16	0	2.89	203.44
cephalexin	15	0	3.26	112.73
cefpirome	10	0	2.67	153.92
ceftazidime	9	–1	2.42	191.22

^aAcidity of prodrug moiety.

For the functional group characterization of the chosen drug molecules, structures were analyzed, and functional groups were established and characterized by standards of organic chemistry.¹² The groups were defined based on pharmacologic and/or chemical significance in a manner to maximize the probability of detection of relevant correlations. In some cases, this led to the establishment of tiers of groups. For example, the unique functional groups pyrrolidinium and pyridinium are each also contained within the broader functional group category of quaternary amines. A complete description of the process by which groups and tiers were established, and a list of the functional groups of cephalosporins analyzed in this study, are available in the Supporting Information. Protein binding data was obtained from the Lexicomp drug monograph for each of the agents included in the analysis.¹³ Protein binding percentages are determined by multiple methodologies, such as equilibrium dialysis assays to determine fraction unbound, radiolabels, micro-extraction, micro-ultrafiltation, or fluorescent tag methods.¹⁵ As the data provided by the database is a compilation from multiple sources, no distinctions were made based on method of protein binding analysis. Statistical analysis was conducted by point-biserial correlation at r = 0.20 and α = 0.05 to identify associations between the presence of functional groups and the percentage of protein binding. Positive correlation was defined as the presence of a group associated with increased protein binding, while negative correlation was defined as the presence of a group associated with decreased protein binding.

Results

A total of 24 cephalosporins and 40 functional groups were analyzed. Four functional groups were associated with protein binding: tetrazole, pyridinium, primary amine, and quaternary amine. Charged (anionic or dianionic) cephalosporins were positively associated with protein binding (r = 0.59, p < 0.01). Results of analysis are summarized in Table 2.

Table 2. Significant Functional Group–Protein Binding Associations.

functional group	cephalosporins containing noted group	correlation coefficient (±r)	p value
tetrazole	cephazolin, cefotetan, cefoperazone	+0.47	<0.05
pyridinium	ceftazidime, ceftaroline	–0.44	<0.05
primary amine	cephalexin, cefadroxil, ceftolozane, cefaclor, cefprozil	–0.43	<0.05
quaternary amine	ceftolozane, ceftazidime, ceftaroline, cefpirome, cefepime	–0.57	<0.01

Discussion

Molecular size, steric environment, and the electronic effects of adjacent functional groups are important determinants of the presence and strength of binding interactions, which confounds both detection of functional groups that may be associated with binding interactions and development of generalizable predictions for their impact on binding.² On the other hand, such associations may be identified by use of available clinical data, which has the advantage of derivation from large sample sizes. When functional groups are present in multiple cephalosporins, intra-class protein binding variability can be linked to specific groups through statistical analysis. Our analysis therefore used available structural and clinical data of cephalosporins to identify correlations to guide future study of their functional groups’ roles in protein binding.

The ionic character of cephalosporins, for both pharmacokinetics and receptor interactions, is a key determinant of cephalosporin protein binding.^2,5,6 Given a positively charged presumptive binding site of albumin, increased binding is predicted to be associated with negatively charged groups or ions, and decreased binding associated with positively charged groups or ions. Study results were consistent with these expectations; functional groups that were associated with decreased binding are positively charged at physiological pH, and negatively charged drugs were associated with increased binding relative to net neutral drugs. These results are useful for validation of the methodology of this study and provide a first-order approximation of effect size on the basis of comparison to this established binding paradigm (negative charge association r = 0.59).

Correlation of protein binding with the presence of a tetrazole group is interesting in that the group is not negatively charged; as an acidic carboxylic acid isostere the group is usually anionic. However, the acidic proton-containing nitrogen of tetrazole (position 1) is either alkyl substituted or the site of attachment to the drug molecule for each of the three cephalosporins that feature the group and therefore cannot be deprotonated. Notably, this group is contained within the R₁ substituent of cefazolin but is located on the R₂ substituents of cefotetan and cefoperazone. However, cefazolin contains thiadiazole (1,3,4) at the analogous position within its R₂ group. The cefotetan and cefoperazone R₂ tetrazoles, as well as the cefazolin thiadiazole, are all attached at the C3 position of the cephem core via a thiomethylene group. Ceftriaxone is the only other cephalosporin that features a similar group (1,2,4-triazine-5,6-dione) at the same position with the same linkage, and it demonstrates extensive protein binding (95%). Additional study, such as in silico molecular docking analysis, is necessary to clarify the roles of the heterocyclic substituents in binding albumin. These cephalosporins and the noted groups are shown in Figure 3.

Figure 3.

Cephalosporins containing tetrazole with similar groups at the R₂ position. Tetrazole groups blue, related groups at R₂ position red, thiomethylene linkage green.

These results could explain the similarity in protein binding of cefazolin and ceftriaxone, and the role in protein binding of aromatic nitrogen heterocycles featuring a thiomethylene linkage at the C3 position warrants further investigation. Although the 1,2,4-triazine-5,6-dione group of ceftriaxone is predominantly negatively charged at physiological pH, tetrazole is neutral in the absence of an acidic hydrogen, indicating potential cation−π or π–π interactions with albumin, which are common in biological systems and modulated by the presence of heteroatoms.¹⁶⁻¹⁸

Quaternary amines, present within R₂ groups of five cephalosporins (three of which are pyridinium rings), were associated with decreased protein binding, as explained by positive charge repulsion with albumin. In addition to decreased protein binding, these groups are associated with enhanced porin penetration and Gram-negative coverage. When present, primary amines are located within the R₁ group on the carbon alpha to the amide of the pharmacophore. One exception is ceftolozane, which features a terminal primary amine within its large R₂ group. Amines located on aromatic systems, such the 2-aminothiazole group present in 12 cephalosporins, were not associated with protein binding. This result is consistent with the notion that the positive charge of amines is the key consideration for protein binding, as amines conjugated with aromatic systems are not sufficiently basic to be protonated at physiological pH in most tissues.¹⁹

An important limitation to this study is the low frequency of occurrence of some functional groups. For example, of 40 functional groups characterized, 15 were present in only one cephalosporin. This maintains the possibility of actual associations not detected by this analysis (beta error). Therefore, while the groups that were significantly associated are likely to play important roles in protein binding, no conclusions can be obtained for other groups (power of analysis is 0.15 based on 24 drugs studied at r = 0.2 and α = 0.05).²⁰ Additionally, a key characteristic of this study is that it consists of analysis of functional groups rather than QSAR or computational study of binding interactions or empirical study of crystal structures of the drugs bound with human serum albumin. This is a limitation in that the analysis does not provide definitive protein binding information. However, as a hypothesis-generating study, these results indicate that future analyses of cephalosporin protein binding by these means should target neutral or negatively charged nitrogen heterocycle substituents at the C3 position.

Conclusion

Charged functional groups and the net charge of cephalosporins are key determinants of protein binding. Positively charged functional groups are associated with decreased protein binding, and a net negative charge is associated with increased protein binding. The presence of a neutral tetrazole group is associated with increased protein binding. Structural analysis of the three highly bound, tetrazole-containing cephalosporins (cefazolin, cefotetan, and cefoperazone), along with ceftriaxone, which is highly bound and features analogous R₂ functionality despite the absence of tetrazole, indicates a need for further study of the role of thiomethylene-linked neutral or negatively charged aromatic nitrogen heterocycles at the C3 position in protein binding.

Glossary

Abbreviations

C3

carbon 3 of cephem core, location of R₂ group

C7

carbon 7 of cephem core, location of R₁ group

FDA

U.S. Food and Drug Administration

EDC

Established Drug Class

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.2c00168.

• Structures of all cephalosporins included in analysis and functional groups of each cephalosporin (PDF)

The authors declare no competing financial interest.