Nanoscale mapping of nucleolin–aptamer interactions on lung cancer cells reveals binding affinity hierarchy and spatial heterogeneity

Abstract

Nucleolin, a protein overexpressed on the surface of cancer cells, has emerged as a promising therapeutic target due to its high affinity interactions with aptamers. This study localized nucleolin on lung cancer and normal cells at single-molecule resolution using the single molecule recognition imaging mode of Atomic Force Microscopy (AFM) with three aptamers: 9FU-AS1411, AS1411, and CRO. The results revealed abundant nucleolin expression on lung cancer cells, while minimal levels were detected on normal cells. The binding affinities and interaction dynamics of these aptamers were systematically evaluated. Flow cytometry and AFM-based force spectroscopy demonstrated that 9FU-AS1411 exhibited the strongest unbinding forces (piconewton level) and higher dissociation activation energy compared to AS1411, indicating enhanced complex stability. In contrast, CRO showed negligible binding, confirming its lack of specificity. Further analysis via Kelvin Probe Force Microscopy (KPFM) revealed distinct surface potential decrements after aptamer interactions: 24.4 mV (9FU-AS1411), 11.7 mV (AS1411), and 2.5 mV (CRO), correlating with their binding strengths. These findings quantitatively rank aptamer affinity as 9FU-AS1411 > AS1411 >> CRO, supported by molecular-level mechanistic insights into electrostatic and structural interactions. This work pioneers high-resolution spatial mapping of nucleolin–aptamer interactions, offering novel methodologies for studying protein–aptamer binding kinetics and electrical properties at unprecedented precision (0.1 mV resolution). The approaches established here not only advance nucleolin-targeted cancer therapy but also provide a framework for investigating other protein–aptamer systems in biomedical research.

---

1 Introduction

Nucleolin is a kind of phosphoprotein that is distributed in plasma membrane, cytoplasm and nucleolus. Nucleolins participate in many crucial physiological activities, such as cell endocytosis, cell proliferation and apoptosis, the assembly of ribosomal subunits, controlling RNA metabolism and ribosome biogenesis.^1,2 More importantly, nucleolins on the cellular surface are involved in tumorigenesis.³ The overexpression and accumulation of nucleolin on the cell membrane are in close relation with many types of cancers, such as lung cancer. It is considered as a biomarker for the diagnosis of cancers.⁴ Thus, it is a promising therapeutic target for cancer treatment.⁵ Nucleolin contains three distinct regions: the acidic N-terminal domain, four central RNA recognition motifs (RBDs or RRMs), and a C-terminal “tail” called the GAR or RGG domain (rich in glycine, arginine and phenylalanine residues).⁶ Aptamers are most frequently used as tumor recognition molecules for targeting nucleolins on the surfaces of cancer cells.⁷

Aptamers are short RNA or single-stranded DNA molecules (with a length of around 20–80 bases) with specific three dimensional structures that can bind target proteins with high selectivity and affinity.⁸ Owing to the lower cost, smaller size, and easier modification compared to antibodies, they have become more ideal recognition candidates for diagnostic and therapeutic agents, biosensing probes and targeted drug delivery systems.⁹ Among them, AS1411 is a guanine-rich (G), 26 base oligonucleotide aptamer capable of binding the external domain of nucleolin.⁷ AS1411 is the first tested and advanced aptamer in cancer therapy and can function as a recognition probe to detect the cell surface nucleolins overexpressed in cancer cells.¹⁰ Watson-Crick and Hoogsteen hydrogen bonds can maintain complex nucleic acid structures under different physicochemical conditions (Fig. 1A).¹¹ The G-quadruplex structure formed by AS1411 is the core basis for its function. Each G-quadruplex is formed by four G bases creating a planar tetrad, with multiple tetrads stacked layer by layer to form a stable three-dimensional structure. The aptamer 9FU-AS1411 is obtained replacing nine thymines (T) in AS1411 with fluorouracil (FU), without changing the sequences of nucleotide bases. 9FU-AS1411 and AS1411 reveal potential use as targeting agents for imaging and theranostic applications of cancers and have the potential to improve cancer diagnoses and treatment.¹²

Fig. 1 The structures of three types of aptamers. (A, B and D) the structures of aptamers AS1411, 9FU-AS1411 and CRO, respectively; (C) the structures of T and FU; (E) the structures of aptamers modified with amino groups.

But till now, localization of nucleolins on lung cancer cell surfaces has not been performed at the single molecule level under physiological conditions, and the interaction forces and kinetics between nucleolins and aptamers on lung cancer cells have not been studied at a very high force resolution level. More importantly, the electrical interactions (such as surface potential) between nucleolins and aptamers have not been investigated. Biomolecules possess distinct charges and consequently exhibit distinct surface potentials. The interactions between biomolecules are in fact electrical interactions and therefore are determined by their electrical properties. Surface potentials and electrostatic interactions in biological systems are key elements of cellular functions, for instance protein folding and assembly.¹³ When membrane proteins bind reagents and medicines, their surface potentials will change, and their conformations will change. This will further in turn influence the related cellular activities. Compared with the physical and chemical properties, surface potential reflects the more essential nature of biomolecules. Thus, surface potential is an important indicator of cellular functions and activities, and measurement of changes in surface potential is a prospective approach for the label free detection of target biomolecules.¹³ But the surface potentials of the interaction processes between nucleolins and aptamers have not been studied so far.

Atomic force microscopy (AFM) offers a powerful platform to address these gaps. Capable of imaging biological samples at nanometer resolution, AFM also enables single molecule force spectroscopy (SMFS) to probe interaction forces and kinetics with piconewton sensitivity.¹⁴ Additionally, the Single Molecule Recognition Imaging (SMRI) mode of AFM allows simultaneous topographic and receptor mapping, facilitating the localization of specific molecules on complex cellular surfaces.^15,16 Furthermore, the electrical modes of AFM, such as Kelvin Probe Force Microscopy (KPFM), can detect surface potential changes with exceptional resolution (0.1 mV), providing insights into electrostatic interactions during binding events.^17–19 KPFM has been successfully employed to study DNA properties, protein-ligand binding, and dynamic molecular processes, demonstrating its potential for label free detection of biomolecules.^20,21

This study leverages the multifunctional capabilities of AFM to achieve high resolution spatial, mechanical, and electrical characterization of nucleolin and aptamer interactions on lung cancer cells. SMRI was employed to localize nucleolin at the single molecule level, SMFS was used to quantify binding forces and kinetics, and KPFM was applied to measure associated surface potential changes. Through comparative analysis of three aptamers (9FU-AS1411, AS1411, and the control CRO), a clear affinity hierarchy was established and mechanistic insights into their interactions with nucleolin were provided. The integrated approach not only advances the understanding of nucleolin targeted binding, but also establishes a versatile framework for investigating protein and aptamer systems across biomedical applications.

2 Results and discussion

2.1 The structures of the three types of aptamers

The structure of AS1411 is shown in Fig. 1A. There are multiple intermolecular interactions in AS1411. First, there are Hoogsteen hydrogen bonds. In each G-tetrad, four G bases form planar structures via N1–H⋯N7 and O6⋯N7 hydrogen bonds (distinct from the Watson-Crick hydrogen bonds in double-stranded DNA). Second, π–π stacking occurs. Hydrophobic stacking forces between tetrad planes enhance overall stability, while hydrogen bonds or van der Waals forces in loop regions (TTA or GT loops) further consolidate the conformation. There is a compact stereoconformation in AS1411. The G-quadruplex forms a columnar structure with stacked planar tetrads, approximately 2 nm in diameter, with height dependent on the number of tetrad layers. The structure contains grooves of varying widths (major and minor grooves), exposing specific functional groups (amino and carbonyl groups of G bases) that serve as recognition sites for protein binding. Flexible loops connecting tetrads (for instance TTA or GT loops in AS1411) can participate in target binding (such as interaction between nucleolin's binding domain and loop regions). Overhangs of terminal G bases facilitate intermolecular interactions. These functions are closely related with their properties. The three-dimensional conformation of the G-quadruplex provides a unique and highly specific target recognition binding interface for nucleolin, with chemical groups in loops and grooves forming hydrogen bonds, hydrophobic interactions, or ionic bonds with target amino acid residues. AS1411 specifically binds to the C-terminal domain of nucleolin via its G-quadruplex structure, blocking its pro-cancer functions (such as cell proliferation and angiogenesis). The compact G-quadruplex structure resists digestion by nucleases (such as DNase), because the enzyme has limited access to internal phosphodiester bonds, significantly prolonging its half-life in vivo and making it suitable as a drug carrier or therapeutic molecule. Meanwhile, AS1411 exhibits conformational dynamics and adaptability. Despite its stability, the G-quadruplex can undergo local conformational changes (e.g., loop unfolding or minor tetrad distortion) during target binding, enabling “induced fit”-type interactions that enhance affinity and specificity. The G-quadruplex structure of AS1411 achieves high affinity binding and functional regulation of targets like nucleolin through sequence specific folding, cation-mediated stabilization, and three-dimensional conformational diversity. Its key features can be summarized as stability via G-base stacking and cation coordination, a unique 3D recognition interface, nuclease resistance, and dynamic adaptability—advantages that make it an ideal molecular template for nucleic acid drug design. Short RNA aptamers and RNA recognition motifs also have significant potential in selectively binding target proteins for biomarker detection. These nucleic acid based recognition elements offer high affinity and specificity, enabling precise identification of disease related proteins even in complex cellular environments. The integration of such molecular tools with advanced sensing platforms provides a powerful approach for transforming protein detection. This strategy not only enhances detection accuracy but also paves the way for novel diagnostic applications in studying various pathological conditions.^22,23

When nine T of AS1411 are replaced with their bioisostere FU, the modified aptamer 9FU-AS1411 can be obtained (Fig. 1B). 9FU-AS1411 preserves the G-quadruplex structure of AS1411, forming stable tetrad stacks via Hoogsteen hydrogen bonds. 5-F in FU has strong electron withdrawing ability, compared with the 5-methyl group in T (Fig. 1C), which enhances the cyclic Hoogsteen-like hydrogen bond structures between four G and enhances the ability to form a G-quadruplex. This bioisosteric replacement strategy preserves the G-quadruplex core structure of AS1411 while improving nuclease resistance through the hydrophobic effects of fluorine atoms. The 5-FU modification site is typically linked to the 5′-end or loop region (e.g., TTA loop) of AS1411, avoiding disruption of the G-quadruplex core conformation. The introduction of 5-FU may enhance the local conformational flexibility of the aptamer. 5-FU modification has minimal impact on the G-quadruplex stability of AS1411, allowing it to maintain targeting ability under physiological conditions.

The aptamer CRO is obtained by replacing all the G in AS1411 with C (cytosines), and there are no cyclic Hoogsteen-like hydrogen bond structures between four C (Fig. 1D). The stability and affinity for nucleolin are expected to be very low, and CRO serves as the control aptamer.

Then, amino groups are added to the three types of aptamers (Fig. 1E); therefore, the three types of aptamers can be functionalized on AFM probes.

2.2 The binding affinities of aptamers with nucleolins

In this work, three types of lung cancer cells and three types of normal cells were used. The three types of lung cancer cells are A549, NCI-H157 and NCI-H1299. One type of normal cell is from the lung, BEAS-2B (human normal lung epithelial cells), and two types of normal cells are from other organs, HSF (human skin fibroblast) and HUVEC (human umbilical vein endothelial cells).

In order to verify that there are nucleolins on the surfaces of lung cancer cells, fluorescence experiments were performed on three types of lung cancer cells and one type of normal lung cell (Fig. S1 in the SI). There are plenty of nucleolins on the three types of lung cancer cells. But there are very few nucleolins on the normal cells (BEAS-2B). The western blot experiments also indicate that there are more nucleolins on lung cancer cells than on normal cells (Fig. S2 in the SI).

Then, the binding affinity and selectivity of the three types of aptamers with lung cancer cells (Fig. 2) and normal cells (Fig. S3 in the SI) over time (0 h–12 h) were measured by flow cytometry. Fig. 2 displays representative raw flow cytometry data. The peaks in Fig. 2 represent the fluorescence intensity distribution of cell populations incubated with FITC labeled aptamers during time lapse. A shift of the peak to the right indicates higher fluorescence intensity, which corresponds to a greater amount of aptamer bound to the cells. The quantitative analysis of this shift, expressed as the percentage of cells that were fluorescently positive (the binding rate), was used to determine affinity and selectivity, and these data are comprehensively detailed in the accompanying Tables 1–3. For 9FU-AS1411 and AS1411, their binding rates with cancer cells are much greater than that with normal cells (Tables 1 and 2), which can contribute to more nucleolins on cancer cells than normal cells, and 9FU-AS1411 and AS1411 can interact with nucleolins specifically. High binding in cancer cells aligns with nucleolin's role as a receptor for AS1411. Overexpression in tumors enhances drug internalization. Prolonged exposure increases binding, likely due to nucleolin recycling or sustained drug retention in cancer cells. For all three types of cancer cell lines, the binding rates of 9FU-AS1411 are larger than those of AS1411, which indicates that the binding affinity and selectivity of 9FU-AS1411 with nucleolins are better than those of AS1411. For CRO, the binding rates with all the cancer and normal cells are very low (Table 3), which indicates that CRO has minimal binding affinity and selectivity with nucleolins. All these demonstrate that the binding affinity and selectivity of aptamers with nucleolins are: 9FU-AS1411 > AS1411 >> CRO.

Fig. 2 The binding affinity and selectivity of three types of aptamers with three different lung cancer cells during time lapse. Representative flow cytometry histograms of (A) A549, (B) NCI-H157, and (C) NCI-H1299 cells incubated with FITC labeled aptamers.

Table 1 Cell binding rate (%) with 9FU-AS1411

	0.5 h	1.5 h	3 h	6 h	12 h
A549	11.2	21.2	30.5	57.6	61.4
NCI-H157	29.5	37.0	38.2	50.5	70.3
NCI-H1299	43.9	44.4	61.6	70.3	84.8
HSF	3.48	4.03	4.89	5.1	6.54
HUVEC	9.72	10.6	12.8	13.5	16.4
BEAS-2B	1.45	1.79	1.73	2.81	1.95

---

Table 2 Cell binding rate (%) with AS1411

	0.5 h	1.5 h	3 h	6 h	12 h
A549	9.04	11.7	19.3	22.8	32.6
NCI-H157	27.8	29.8	33.5	36.8	56.5
NCI-H1299	28	36.6	40.4	41.1	41.3
HSF	4.13	4.01	3.85	4.4	4.26
HUVEC	7.48	8.16	10.7	13.3	13.8
BEAS-2B	2.03	2.00	1.96	2.64	3.05

---

Table 3 Cell binding rate (%) with CRO

	0.5 h	1.5 h	3 h	6 h	12 h
A549	6.78	5.46	6.91	8.25	6.15
NCI-H157	6.12	4.73	4.27	4.39	6.35
NCI-H1299	1.26	2.43	2.84	3.59	4.74
HSF	3.16	3.26	3.05	3.19	3.42
HUVEC	7.85	9.12	8.75	9.04	9.91
BEAS-2B	1.11	1.00	0.90	1.14	1.19

---

2.3 Localization of nucleolins on the surfaces of lung cancer cells and normal cells by 9FU-AS1411 modified tips

As the resolution of the fluorescence images is limited, the distributions of nucleolins on the surfaces of lung cancer cells and normal cells were investigated using the SMRI mode of AFM at single molecular resolution. In this mode, the tips were modified with an aptamer via the flexible and heterobifunctional polyethylene glycol (PEG) crosslinker as depicted in Fig. 3A. One aptamer molecule can only connect one PEG crosslinker. The nonlinear stretching characteristics of PEG make it possible to distinguish the specific events from the nonspecific ones. As PEG is inert in physical and chemical properties, the aptamer functionalized on the tips can reorientate freely and rapidly when the tip is scanning the sample surface. Meanwhile, the PEG tethered aptamer prevents the tip from crashing.²⁴ The amplitude of the cantilever is set to be smaller than the stretched length of the PEG crosslinker.¹⁶ When the sites of nucleolin on the cellular membranes are scanned by this tip, the crosslinker will be stretched during the retraction processes of the cantilever. There will be energy loss, which can reduce the top peak of the oscillations. Thus, the recognition signals can be detected.²⁵ The recognition processes have been proved to be highly specific, reproducible and efficient.¹⁶

Fig. 3 Localization of nucleolins by aptamer 9FU-AS1411 modified tips. (A) The schematic diagram of the AFM tip modified with the aptamer; (B and C) the topography and corresponding recognition images of A549 cells; (D) the recognition image after blocking; (E and F) the recognition images acquired on the surfaces of NCI-H157 and BEAS-2B cells by 9FU-AS1411-NH₂ modified tips, respectively; (G and H) the typical force curves on A549 cells before and after blocking, respectively; (I and J) the histograms of the distributions of the unbinding forces obtained on the surfaces of A549 cells before and after blocking, respectively.

Firstly, nucleolins on the surfaces of lung cancer and normal cells were localized by 9FU-AS1411 modified tips. The topography and corresponding recognition images captured on A549 cells are shown in Fig. 3B and C, respectively. The dark spots in Fig. 3C represent the recognition signals that were irregularly distributed on A549 cell membranes. The recognition signals occupy an area percentage of (9.3 ± 0.7)% (N = 6) of the cell membranes. Therefore, the quantitative information of the distributions and amounts of nucleolin on A549 cell membranes was directly obtained. In order to indicate the distributions of nucleolins on A549 cell membranes more clearly, the recognition signals were superimposed onto the topographic image (the green areas in Fig. 3B). Blocking experiments were carried out by addition of the free aptamer 9FU-AS1411 into the AFM sample cell, for the sake of verification of the specificity of the recognition processes. The binding sites were occupied by the free aptamer and the recognition signals disappeared (Fig. 3D). In order to further verify the specificity of the interactions between the aptamer 9FU-AS1411 and nucleolins, control experiments were performed using bare tips or only PEG modified tips (Fig. S4 in the SI). There are no recognition signals in these images. All these confirm that the recognition signals are indeed from the specific interactions between the nucleolins on the surfaces of A549 cells and aptamers modified on the tips.

Similar experiments were carried out on the surfaces of other types of lung cancer cells (NCI-H157, Fig. 3E) and normal cells (BEAS-2B, Fig. 3F) by aptamer 9FU-AS1411 modified tips. There are plenty of recognition signals on NCI-H157 cells which represent the nucleolin sites (Fig. 3E). There are almost no recognition signals on BEAS-2B cells, which means that there are almost no nucleolins on BEAS-2B cells (Fig. 3F). All these demonstrate high expression of nucleolins on different types of lung cancer cells, and there are almost no nucleolins on normal cells.

Then, the interactions between the nucleolins and the 9FU-AS1411 were studied using the SMFS mode of AFM. In this mode, the aptamer 9FU-AS1411 was attached to the AFM tips, as shown in Fig. 3A. When the tip approaches and withdraws from the surface of cells, the interaction forces between the aptamer modified on the tip and the nucleolins on the cell membranes can be detected and recorded as force curves. Thousands of force curves were recorded at various positions on different cells. The typical force curve acquired on the surface of A549 cells with one unbinding force event is shown in Fig. 3G. The approaching and withdrawn processes are depicted as black and red curves, respectively. The unbinding forces range from 36 to 121 pN at a loading rate of 0.47 nN s⁻¹, with the maximum distribution at 69.4 ± 16.7 pN (Fig. 3I). The binding probability, the overall force curves divided by those with the specific unbinding force event, is 21.8%. After blocking by the addition of free 9FU-AS1411, the specific unbinding force event disappeared (Fig. 3H), and the binding probability has dramatically decreased to 2.1% (Fig. 3J). There are no unbinding force events in the force curves acquired with bare tips or PEG modified tips (Fig. S5 in the SI). SMFS control experiments were also carried out on the surfaces of normal cells (Fig. S6 in the SI). There are no specific unbinding force events, which further indicates that there are almost no nucleolins on normal cells (either from the lung or from other organs). All these confirmed that the unbinding forces between nucleolins on the surface of A549 cells and 9FU-AS1411 on the tips are detected specifically and efficiently.

The combination of AFM and other detection methods at nanometer or single molecule resolution (such as nanopore sensing) offers a powerful platform for investigating protein folding and self-assembly at the nanoscale. AFM provides high resolution visualization of morphological details such as fibril twisting and polymorphism, as demonstrated in the heparin mediated Tau enantiomer assembly. Meanwhile, nanopore technology enables real time, label free detection of dynamic intermediates and aggregation states at the single molecule level. This dual approach allows researchers to correlate structural changes with kinetic behavior, offering deeper insights into chiral specific protein interactions and aggregation pathways. Such multimodal nanoscale analysis significantly enhances our understanding of complex biomolecular processes underlying related diseases.²⁶

2.4 Localization of nucleolins on the surfaces of lung cancer cells and normal cells by AS1411 modified tips

Then, the nucleolins on the surfaces of lung cancer cells were localized by AS1411 modified tips. The morphology and corresponding recognition images of A549 cells are depicted in Fig. 4A and B, respectively. There are plenty of recognition signals of nucleolins (Fig. 4B). In order to reveal the distributions of nucleolins on the cellular surface more clearly, the recognition dots were imposed on the topography (Fig. 4A). The recognition sites occupy an area percentage of (8.9 ± 0.6)% (N = 6), which is almost the same as the value from 9FU-AS1411 modified tips, which indicates that AS1411 can also recognize most of the nucleolins. After blocking, the recognition signals disappeared (Fig. 4C), which indicates that the dark spots in Fig. 4B are indeed the nucleolin recognition signals.

Fig. 4 Single molecule recognition of nucleolin on the surface of A549 cells by aptamer AS1411 modified tips. (A and B) The topography and corresponding recognition images of A549 cells, respectively; (C) the recognition image after blocking; (D and E) the typical force curves before and after blocking, respectively; (F and G) the histograms of distributions of unbinding forces before and after blocking, respectively; (H) the plots of the relationship of the most probable unbinding force versus the natural logarithm of the loading rate (pN s⁻¹) of the tip on the surfaces of A549 cells. The green squares and red dots are the data captured by 9FU-AS1411 and AS1411 modified tips, respectively.

Then, the interaction forces between nucleolins and AS1411 were measured by SMFS. The typical force curves before and after blocking are shown in Fig. 4D and E, respectively. There is a specific unbinding force event in the force curve (Fig. 4D). After blocking, the specific unbinding force event disappeared. Before blocking, the binding probability is 18.0%. The unbinding force is in the range of 26–74 pN with the maximum distributions at (50.4 ± 12.7) pN at a loading rate of 0.47 nN s⁻¹ (Fig. 4F). After blocking, the binding probability has decreased to 2.8% (Fig. 4G).

The unbinding forces not only depend on the interactions between the aptamers and nucleolins, but also depend on the loading rates of the AFM tip. According to the single barrier model, the relationship between the unbinding forces and loading rates follows eqn (1):

where F_u is the unbinding force; k_B is the Boltzmann constant; T is the thermodynamic temperature; x_β is the separation energy barrier from the equilibrium position; r is the loading rate of the tip; k_off is the dissociation rate constant at zero force.²⁷

The plots of the dependence of the unbinding forces versus loading rates are shown in Fig. 4H. The blue dots and red squares are the data captured by 9FU-AS1411 and AS1411 modified tips, respectively. From the fitting curves, the following parameters are inferred: for 9FU-AS1411, x_β = 0.27 nm and k_off = 0.24 s⁻¹; for AS1411, x_β = 0.37 nm and k_off = 0.30 s⁻¹. The dissociation rate constant for 9FU-AS1411 is lower than that for AS1411, which reveals that the nucleolin-(9FU-AS1411) complex is more stable than the nucleolin-AS1411 complex.

Based on transition state theory, the relationship between activation energy ΔE and the dissociation rate constant k_off follows eqn (2).

The difference values of dissociation activation energy Δ(ΔE) of the nucleolin–aptamer complex system follow eqn (3):²⁸

Therefore, it can be calculated that Δ(ΔE) of nucleolin-(9FU-AS1411) and nucleolin-AS1411 is 0.22 k_BT. It indicates that the dissociation activation energy of nucleolin-(9FU-AS1411) is higher than that of nucleolin-AS1411, which further reveals that the nucleolin-(9FU-AS1411) complex is more stable and harder to dissociate than nucleolin-AS1411. These are difficult (or even impossible) to be studied by other approaches. Though nucleolins can be recognized by 9FU-AS1411 and AS1411, the affinities between nucleolin-(9FU-AS1411) and nucleolin-AS1411 are different. SMFS has revealed that the binding force between 9FU-AS1411 and nucleolin is stronger than that between AS1411 and nucleolin. This phenomenon can be attributed to the following key molecular mechanisms and structural optimization. Firstly, there are enhanced intermolecular interactions via 5-FU modification and the synergy of hydrogen bonds and hydrophobic interactions. The pyrimidine ring of 5-FU forms additional hydrogen bonds with amino acid residues on nucleolin. For example, the fluorine atom of 5-FU acts as a hydrogen bond acceptor, stabilizing polar interactions with the positively charged guanidinium groups (e.g., Arg side chains) of nucleolin. Additionally, the hydrophobic pyrimidine ring of 5-FU may embed into hydrophobic pockets of nucleolin, further enhancing binding stability via van der Waals forces. Meanwhile, there are cooperative effects with the G-quadruplex. The G-quadruplex structure of AS1411 is stabilized by Hoogsteen hydrogen bonds. 5-FU modification may reinforce this stability through π–π stacking with the guanine bases of the G-quadruplex. For instance, the pyrimidine ring of 5-FU stacks with the planar G bases, enhancing the overall rigidity of the aptamer and reducing conformational fluctuations during nucleolin binding, thus increasing binding force. Secondly, linker design optimizes binding kinetics, and flexible linkers promote “induced fit”. 5-FU in 9FU-AS1411 is typically linked to the 5′-end or loop region of AS1411 via a flexible linker. This design allows 5-FU to dynamically adjust its position through conformational (swinging) changes during binding, enabling a more precise “induced fit” with nucleolin's binding sites. The flexibility of the linker may help 5-FU access hidden sites on nucleolin, increasing contact area and binding strength. There is a reduced entropic penalty. Flexible linkers minimize conformational restrictions during aptamer binding, reducing entropy loss (ΔS) and improving binding free energy (ΔG). In AFM experiments, this manifests as a higher rupture force, as a greater mechanical force is required to disrupt the dynamically optimized binding conformation. Thirdly, the 5-FU modification enhances binding by stabilizing the G-quadruplex and promoting a preorganized conformation that is favorable for nucleolin recognition. The introduction of 5-FU may alter the local conformation of AS1411 via steric hindrance. For example, modification in the TTA loop region may enhance loop flexibility, promoting tighter interactions with nucleolin's binding pockets. 5-FU modification may stabilize a specific conformation (e.g., parallel G-quadruplex), optimizing binding affinity to nucleolin. Fourthly, there are multivalent binding and synergistic effects. As a bifunctional molecule, 9FU-AS1411 combines the nucleolin-targeting ability of AS1411 with the chemotherapeutic activity of 5-FU. This synergy enables multivalent binding modes, such as multiple binding sites on nucleolin (e.g., RBD1 and RBD2 domains) simultaneously interacting with different regions of AS1411, forming “multi-point anchoring”. Additional interactions between 5-FU and nucleolin further stabilize the complex, leading to a higher rupture force in AFM measurements. AFM measures the maximum rupture force of single-molecule interactions, and binding of 9FU-AS1411 may involve dynamic binding-dissociation cycles. The presence of 5-FU prolongs the binding duration (reduces the dissociation rate constant k_off), resulting in a higher force value in force spectra.

Molecular dynamics simulations reveal that the fluorine atom of 5-FU in 9FU-AS1411 forms a hydrogen bond with Arg123 of nucleolin, and the flexible linker allows the aptamer to more fully occupy nucleolin's hydrophobic pockets during binding.⁷

9FU-AS1411 significantly enhances binding force to nucleolin through a dual strategy of chemical modification and structural optimization, retaining AS1411's targeting specificity while providing a molecular basis for efficient tumor delivery. Future research could further validate these mechanisms using site-directed mutagenesis and cryo-electron microscopy to resolve binding conformations.

The measured difference in dissociation activation energy between the nucleolin-9FU-AS1411 and nucleolin-AS1411 complexes has profound and direct implications for cellular targeting and therapeutic potential. The related discussion has been expanded in the revised manuscript to clarify these points.

The observed stability difference directly influences complex lifetime. The lower dissociation rate constant means that the 9FU-AS1411-nucleolin complex remains associated longer. This extended lifetime enhances the probability of successful internalization via nucleolin mediated endocytosis, a dynamic cellular process that requires a stable ligand–receptor interaction. This mechanistic advantage is supported by the flow cytometry results, which consistently showed higher binding rates for 9FU-AS1411 across all tested cancer cell lines and time points.

From a therapeutic perspective, the improved binding stability translates to more effective cellular targeting. A longer residence time increases the efficiency of drug delivery if the aptamer is used as a targeting agent, as it allows more cargo to be internalized before dissociation occurs. Furthermore, the tighter binding makes 9FU-AS1411 a more effective antagonist of nucleolin mediated signaling pathways that support tumor growth. The enhanced mechanical stability also suggests better performance under physiological conditions, such as blood flow and shear stress, which would promote greater accumulation and retention at the tumor site in vivo.

The measured energetic difference represents a meaningful improvement in the context of drug development. Such enhancements in affinity and complex stability are often the key factors that differentiate a candidate therapeutic. Thus, the increased dissociation activation energy for 9FU-AS1411 is not merely a biophysical observation but an indicator of its superior potential for targeted cancer therapy and diagnostics.

2.5 Localization of nucleolins on the surfaces of lung cancer cells by CRO modified tips

Then, localization of nucleolins on the surfaces of lung cancer cells was performed by CRO modified tips. The topography and corresponding recognition images of A549 cells are shown in Fig. 5A and B, respectively. There are no recognition signals in the recognition image (Fig. 5B). The interaction force between nucleolins and CRO was detected by SMFS, and there is no specific unbinding force event in the typical force curve (Fig. 5C). All these indicate that there are no specific interactions between CRO and nucleolins, which is in accordance with the fact that CRO is not the specific aptamer against nucleolins. From the force investigations of three types of aptamers with nucleolins, the interaction forces between aptamers and nucleolins are 9FU-AS1411 > AS1411 >> CRO.

Fig. 5 Single molecule recognition of nucleolin on the surface of A549 cells by aptamer CRO modified tips. (A and B) The topography and the corresponding recognition images of A549 cells, respectively; (C) the typical force curve obtained by CRO-NH₂ modified tips.

2.6 Investigations of the electrical interactions between aptamers and nucleolins by KPFM

The electrical interactions (surface potential) between aptamers and nucleolins were studied by KPFM (Fig. 6). First, the effects of 9FU-AS1411 on the surface potential of nucleolins were studied by KPFM (Fig. 6A–F). Before and after the addition of 9FU-AS1411, little changes can be seen from the topography images of nucleolin (Fig. 6A and D). But there are obvious changes in CPD images (Fig. 6B and E). From the distributions of CPD, it can be inferred that the maximum distribution of CPD of nucleolin is at 290.4 ± 4.2 mV (Fig. 6C). After the addition of 9FU-AS1411, the maximum distribution of CPD of nucleolin is at 266.0 ± 4.3 mV (Fig. 6F). The CPD has decreased 24.4 mV. In this AFM setup, the CPD = V_sample − V_tip, where V_sample and V_tip are the surface potentials of the sample and tip, respectively. Before and after the addition of aptamers, the surface potential of the tip does not change. Thus, the changes in CPD of the sample correspond to the changes in the surface potential. Therefore, after the addition of 9FU-AS1411, the surface potential of the nucleolin has decreased 24.4 mV.

Fig. 6 The interactions between the three types of aptamers and nucleolins studied by KPFM. (A and D) The topography images of nucleolin before and after the addition of 9FU-AS1411, respectively; (B and E) the CPD images before and after the addition of 9FU-AS1411, respectively; (C and F) the histograms of distributions of CPD before and after the addition of 9FU-AS1411, respectively; (G and J) the topography images of nucleolin before and after the addition of AS1411, respectively; (H and K) the CPD images before and after the addition of AS1411, respectively; (I and L) the histograms of distributions of CPD before and after the addition of AS1411, respectively; (M and P) the topography images of nucleolin before and after the addition of CRO, respectively; (N and Q) the CPD images before and after the addition of CRO, respectively; (O and R) the histograms of distributions of CPD before and after the addition of CRO, respectively.

The effects of AS1411 on the surface potential of nucleolin have been studied following the same way (Fig. 6G–L). After the addition of AS1411, the surface potential of nucleolin has decreased 11.7 mV (Fig. 6I and L), which is smaller than the value of 9FU-AS1411. This may indicate that the binding affinity and the associated electrostatic interactions of AS1411 for nucleolin are weaker than that of 9FU-AS1411.

The effects of CRO on the surface potential of nucleolin have been studied by KPFM (Fig. 6M–R). After the addition of CRO, only slight changes in the surface potential of the nucleolin are observed (about 2.5 mV, Fig. 6O and R), which is much smaller than the values of 9FU-AS1411 and AS1411. This is likely because CRO is not the specific aptamer against nucleolin and there are no specific interactions between CRO and nucleolin. The CRO experiments indicate that not all aptamers or reagents added will change the surface potential of nucleolins, and only those aptamers that can interact with nucleolins will change the surface potential of nucleolins.

Therefore, from the KPFM study, the effects of aptamers on nucleolins are: 9FU-AS1411 > AS1411 >> CRO, which are consistent with the results of the flow cytometry and force studies.

Former KPFM research has proved that different proteins can be distinguished based on their different isoelectric points. These studies have demonstrated that KPFM can be applied in the label free detection of biological binding events. There are several advantages in the detection of surface potential of biomolecules by KPFM. For instance, it is a non-contact and non-destructive approach and can be performed under ambient conditions, and the obtained signals remain fidelity.²⁹ Biological molecules are usually in a charged state, which is reflected by the isoelectric point of the molecule. It is hypothesized that the isoelectric point directly correlates with the surface potential. KPFM measures the local surface potential of nucleolins, and the surface potential can be considered as the energy surface for the electron. Thus, areas with low surface potential can be regarded as areas with large local electronegativity. The isoelectric point stands for the relative proton affinity of a molecule and can roughly be regarded as the opposite of the electronegativity. Therefore, low surface potential correlates with high electronegativity, and in turn correlates with a low isoelectric point.²⁹ In this work, after binding with 9FU-AS1411 and AS1411, the surface potential of nucleolins decreases, the electronegativity of nucleolins increases, and the isoelectric point of nucleolins decreases. All these demonstrate that KPFM is able to investigate the surface potential of biomolecules interacting with ligands in a label free manner. Furthermore, measured surface potentials during biomolecular interactions enable quantitative descriptions of the ability of proteins to interact with small ligands (for instance the aptamers). KPFM allows the precise recognition of single molecule interactions, which opens a new avenue for the design and development of novel molecular therapeutics.

The mechanism by which the surface potential decreases after 9FU-AS1411 and AS1411 bind to nucleolin involves two main aspects: charge neutralization effects and conformational changes.

Firstly, there is charge neutralization, that is to say, interactions between positively charged nucleolin and negatively charged RNA. Nucleolin is a protein rich in arginine and lysine, with an isoelectric point of about 11.9. At physiological pH (7.4), it carries a large number of positive charges. 9FU-AS1411 is an RNA aptamer whose phosphate backbone carries negative charges under physiological conditions. The 9-FU modification may further increase local negative charge density through the electron-withdrawing effect of electronegative fluorine atoms. For instance, when 5-fluorouracil binds to polyoxometalates, its charge density decreases, and characteristic vibration peaks shift to higher frequencies, suggesting that the modification alters the RNA's electron cloud distribution. When 9FU-AS1411 or AS1411 binds to nucleolin, positively charged residues of nucleolin (e.g., Arg and Lys) interact electrostatically with the phosphate groups of RNA, neutralizing part of the negative charges. This direct charge neutralization reduces the net charge on the complex surface, thereby lowering the surface potential. For example, the specific binding of nucleolin to AS1411-G4 may involve hydrogen bonds and ionic bonds between arginine side chains and RNA phosphate groups.

Secondly, there are conformational changes, namely the formation and compactization of the G-quadruplex structure. Binding to nucleolin may induce further compactization of the G-quadruplex. This structural change sequesters more negative charges inside the complex, reducing the exposure of surface charges and lowering the surface potential. A steric hindrance effect may occur. Nucleolin binding may restrict the flexibility of the RNA chain through steric hindrance, limiting the solvent exposure of phosphate groups. For instance, the formation of a protein-RNA complex may fix flexible regions of RNA, reducing dynamic exposure of surface charges.

The core mechanism for the reduction in surface potential after 9FU-AS1411 and AS1411 bind to nucleolin is charge neutralization, where positively charged residues of nucleolin neutralize the negative charges of RNA. Additionally, the formation and compactization of the G-quadruplex structure further reduce the exposure of surface charges. These factors collectively lead to a significant decrease in the surface potential of the complex. This mechanism not only explains experimental observations but also provides a theoretical basis for the design of aptamer-based targeted drugs.

The significantly greater reduction in surface potential when 9FU-AS1411 binds to nucleolin compared to AS1411 alone arises from the synergistic effects of differences in charge properties, enhanced structural stability, and optimized intermolecular interactions.

Firstly, there are differences in charge properties, arising from the RNA backbone and 9FU modification. When 9FU-AS1411 binds to nucleolin, its high density negative charges undergo stronger electrostatic neutralization with the positively charged residues of nucleolin (e.g., arginine and lysine), leading to a more pronounced decrease in surface potential. The fluorine atom in 9-fluorouracil (9FU) has high electronegativity, enhancing local negative charge density on RNA via its electron withdrawing effect. For instance, in the binding of 5-fluorouracil to polyoxometalates, fluorine introduction reduces molecular charge density and shifts characteristic vibration peaks to higher frequencies. This effect may make the phosphate groups of 9FU-AS1411 more accessible to nucleolin's positive charges, further enhancing charge neutralization.

Then, there are optimized intermolecular interactions and differences in aptamer–protein binding modes. Higher affinity means nucleolin molecules bind to 9FU-AS1411 and neutralize more negative charges. (2) Steric Hindrance and Conformational Locking. The larger volume of the 9FU moiety may restrict RNA chain flexibility via steric hindrance more effectively than AS1411, leading to a more compact protein-RNA complex upon binding. This conformational locking reduces dynamic exposure of surface charges, amplifying the surface potential decrease more than AS1411.

The pronounced decrease in surface potential upon (9FU-AS1411)-nucleolin binding results from charge density differences, enhanced structural stability, and optimized binding modes. RNA's higher negative charge density and the electron-withdrawing effect of 9FU provide a stronger basis for electrostatic neutralization, while the thermodynamic stability of RNA G-quadruplexes and metal ion synergy further reduce surface charge exposure. These mechanisms not only explain experimental observations but also offer new optimization directions for RNA aptamer-based targeted drug design.

3 Conclusions

This study successfully employed AFM to achieve nanoscale mapping and quantification of nucleolin–aptamer interactions on lung cancer cells at single-molecule resolution. Key findings demonstrate that nucleolin is abundantly expressed on the surfaces of lung cancer cells (A549, NCI-H157, and NCI-H1299), while being minimally present on normal cells (BEAS-2B, HSF, and HUVEC). Utilizing a multimodal AFM approach, we systematically evaluated three aptamers (9FU-AS1411, AS1411, and CRO) and established a clear affinity hierarchy: 9FU-AS1411 > AS1411 >> CRO. SMRI provided direct spatial localization of nucleolin, revealing its irregular distribution and higher density on cancer cell membranes compared to normal cells. SMFS quantified unbinding forces and kinetics, showing that 9FU-AS1411 exhibits significantly stronger binding and a lower dissociation rate constant than AS1411. This enhanced stability is attributed to optimized intermolecular interactions (e.g., hydrogen bonding and hydrophobic effects) and structural rigidity conferred by fluorouracil (FU) substitution. Conversely, CRO showed negligible binding, confirming its lack of specificity. KPFM further revealed distinct surface potential decrements upon aptamer binding: 24.4 mV (9FU-AS1411), 11.7 mV (AS1411), and only 2.5 mV (CRO). These changes correlate with binding strength and reflect charge neutralization and conformational changes during complex formation. The 0.1 mV resolution of KPFM enabled unprecedented label-free detection of biomolecular electrostatic interactions.

Collectively, this work pioneers high-resolution spatial, mechanical, and electrical profiling of nucleolin–aptamer interactions. The superiority of 9FU-AS1411 highlights its potential for targeted cancer therapy and diagnostics. The integrated AFM methodology establishes a versatile framework for investigating protein–ligand interactions across biomedical research, offering profound insights into binding kinetics, affinity, and electrostatics at the single-molecule level.

Author contributions

L. C. and R. F. performed the fluorescence microscopy and AFM experiments and obtained the data. Q. X. and H. Z. performed the cell culture experiments. W. Z. supervised the whole work, analyzed the data and wrote the original draft. Y. S., W. P. and F. J. reviewed and edited the manuscript. All authors approved the final version of the manuscript.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: contains supporting experimental data, including fluorescence imaging, Western blot analysis, flow cytometry results, and AFM control experiments, along with experimental sections. See DOI: https://doi.org/10.1039/d5ay01088h.

Acknowledgements

This work was supported by the Natural Science Foundation of Jiangxi Province (20212ACB206001 and 20202ZDB01018), the Talent project of Jiangxi Province (jxsq2018106059), and the startup funds of Gannan Medical University (QD202011).

References

1. B. Ferrara, S. Belbekhouche, D. Habert, C. Houppe, B. Vallée, S. Bourgoin-Voillard, J. L. Cohen, I. Cascone and J. Courty, Nanotechnology, 2021, 32, 322001–322027.
2. H. Ginisty, H. Sicard, B. Roger and P. Bouvet, J. Cell Sci., 1999, 112, 761–772.
3. W. Jia, Z. Yao, J. Zhao, Q. Guan and L. Gao, Life Sci., 2017, 186, 1–10.
4. M. Koutsioumpa and E. Papadimitriou, Recent Pat. Anticancer Drug Discov., 2014, 9, 137–152.
5. L. Li, J. Hou, X. Liu, Y. Guo, Y. Wu, L. Zhang and Z. Yang, Biomaterials, 2014, 35, 3840–3850.
6. S. Arumugam, M. C. Miller, J. Maliekal, P. J. Bates, J. O. Trent and A. N. Lane, J. Biomol. NMR, 2010, 47, 79–83.
7. L. Bie, Y. Wang, F. Jiang, Z. Xiao, L. Zhang and J. Wang, Front. Mol. Biosci., 2022, 9, 1025313–1025324.
8. R. Yazdian-Robati, P. Bayat, F. Oroojalian, M. Zargari, M. Ramezani, S. M. Taghdisi and K. Abnous, Int. J. Biol. Macromol., 2020, 155, 1420–1431.
9. P. J. Bates, E. M. Reyes-Reyes, M. T. Malik, E. M. Murphy, M. G. O'Toole and J. O. Trent, Biochim. Biophys. Acta, Gen. Subj., 2017, 1861, 1414–1428.
10. P. J. Bates, D. A. Laber, D. M. Miller, S. D. Thomas and J. O. Trent, Exp. Mol. Pathol., 2009, 86, 151–164.
11. W. Li, J. Li, L. Wang, Y. Wang, Z. Zhang, S. Liu, B. Gong, Y. Wang and L. Wang, Nano Lett., 2025, 25, 1706–1714.
12. V. R. Sharma, S. D. Thomas, D. M. Miller and F. Rezzoug, Cancer Invest., 2018, 36, 475–491.
13. B. Bhushan, and M. L. B. Palacio, in Encyclopedia of Nanotechnology, Springer, Dordrecht, 2016, pp. 1711–1717.
14. L. Wang, Y. Qian, Y. Sun, B. Liu and G. Wei, Rev. Anal. Chem., 2020, 39, 116–129.
15. W. Zhao, S. Liu, M. Cai, H. Xu, J. Jiang and H. Wang, Chem. Commun., 2013, 49, 2980–2982.
16. C. Stroh, H. Wang, R. Bash, B. Ashcroft, J. Nelson, H. Gruber, D. Lohr, S. M. Lindsay and P. Hinterdorfer, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 12503–12507.
17. C. Leung, H. Kinns, B. W. Hoogenboom, S. Howorka and P. Mesquida, Nano Lett., 2009, 9, 2769–2773.
18. H. F. Knapp, P. Mesquida and A. Stemmer, Surf. Interface Anal., 2002, 33, 108–112.
19. I. Lee, J. W. Lee, A. Stubna and E. Greenbaum, J. Phys. Chem. B, 2000, 104, 2439–2443.
20. K. Kwak, S. Yoda and M. Fujihira, Appl. Surf. Sci., 2003, 210, 73–78.
21. J. Park, J. Yang, G. Lee, C. Y. Lee, S. Na, S. W. Lee, S. Haam, Y. M. Huh, D. S. Yoon, K. Eom and T. Kwon, ACS Nano, 2011, 5, 6981–6990.
22. L. Wang, S. Zhou, Y. Wang, Y. Wang, J. Li, X. Chen, D. Zhou, L. Liang, B. Yin, Y. Zhang and L. Wang, Faraday Discuss., 2025, 257, 60–72.
23. L. Zhao, Y. Deng, Y. Wang, S. Zhou, B. Yin, Y. Chen, Y. Wang, J. Li, L. Wang, Y. Lin and L. Wang, Mater. Today Phys., 2024, 46, 101479.
24. G. Kada, F. Kienberger and P. Hinterdorfer, Nano Today, 2008, 3, 12–19.
25. A. Ebner, F. Kienberger, G. Kada, C. M. Stroh, M. Geretschläger, A. S. Kamruzzahan, L. Wildling, W. T. Johnson, B. Ashcroft, J. Nelson, S. M. Lindsay, H. J. Gruber and P. Hinterdorfer, Chemphyschem, 2005, 6, 897–900.
26. W. Chen, Y. Chen, Y. Wang, L. Chen, S. Zhou, D. Li, B. Yin, Z. Yang, L. Wang and H. Wang, Mater. Today Phys., 2024, 42, 101370.
27. E. Evans and K. Ritchie, Biophys. J., 1997, 72, 1541–1555.
28. J. Yu, Y. Jiang, X. Ma, Y. Lin and X. Fang, Chem.–Asian J., 2007, 2, 284–289.
29. A. K. Sinensky and A. M. Belcher, Nat. Nanotechnol., 2007, 2, 653–659.

---

Footnote

† These authors contributed equally to this work.

---