Abstract

   The nucleolus is essential for ribosome biogenesis and stress
   regulation. However, because of its dynamic nature, there is still a
   lack of methods to specifically visualize nucleolar localization in
   living cells and to study dynamic changes in protein interaction
   networks within the cell nucleolus. In this study, we identified and
   engineered a signal peptide sequence, termed nucleolar beacon, which
   exhibits robust nucleolar localization and universal applicability
   across various mammalian cell types. Using this sequence, we
   established nucleolar indicator cell lines and demonstrated their
   practicality in studying nucleolar functions in living cells. In
   addition, by combining the signal peptide with proximity labeling
   technology, we developed an effective approach for capturing the
   nucleolar proteome and successfully identified nucleolar-associated
   proteins. These techniques provide effective and versatile tools for
   investigating nucleolar functions in living cells and offer a potential
   strategy for drug delivery applications.
     __________________________________________________________________

   NoB serves as a reliable tool for visualizing nucleolar dynamics and
   identifying nucleolar-associated proteins.

INTRODUCTION

   The nucleolus, the predominant membrane-less organelle in eukaryotic
   cells, plays a critical role in ribosome biogenesis and serves as a
   central hub for integrating internal and external signals associated
   with growth and stress ([42]1, [43]2). It exhibits a highly organized
   structure and liquid-like properties, which facilitate rapid adaptation
   to cellular homeostasis changes ([44]3). Damage to nucleoli,
   alterations in ribosomal proteins (RPs), and various stressors (such as
   cytotoxic agents, heat shock, nutrient deficiency, ultraviolet
   radiation, and viral infection) can lead to changes in nucleolar
   morphology ([45]4). These changes have been implicated in diverse
   diseases, including neurodegeneration (e.g., Parkinson’s disease),
   cancer (e.g., lung cancer and melanoma), and aging ([46]1, [47]5).
   Clinically, increased nucleolar size or number often signals poor
   prognosis in tumors due to heightened translational demands in rapidly
   dividing cells ([48]2, [49]6, [50]7). Understanding how external
   stimuli and intracellular factors influence the nucleolar dynamic
   alterations is thus critical for elucidating its role in disease.

   Traditional methods for studying nucleolar morphology and function,
   such as chemical staining, ribosomal RNA (rRNA) probes, and RP
   analysis, have yielded important insights but also have limitations
   ([51]8–[52]13). AgNOR staining, for instance, is commonly used to
   visualize the nucleolus and aid in the identification of tumor stage,
   grade, and aggressiveness ([53]10), yet it fails to fully capture
   nucleolar morphology. Recently developed rRNA-selective probes,
   including thioflavin T, 4MPS-TO, and low-energy red luminescent
   alkynylplatinum(II) complexes, offer rapid imaging of the nucleolus in
   live and fixed cells ([54]14–[55]16). Commercially available RNA stains
   such as Nucleolus Bright Red/Green (Dojindo), SYTO RNA Select (Thermo
   Fisher Scientific), and PhenoVue 512 (PerkinElmer) further expand this
   toolkit. However, these probes are constrained by nonspecific rRNA
   binding, poor water solubility, short emission wavelength, low
   stability against photobleaching, high cytotoxicity, elevated cost,
   spectral interference with other common fluorophores, and insufficient
   cellular membrane permeability. Fusion-based reporter systems such as
   green fluorescent protein (GFP)–RPL37 ([56]17), GFP-RPL29 and yellow
   fluorescent protein–RPS2 ([57]18), RPS9-HaloTag ([58]19), FBL-GFP
   ([59]20), upstream binding factor–HaloTag ([60]21), and enhanced GFP
   (EGFP)–DKC1 or eGFP–nucleophosmin 1 (NPM1) ([61]22) play a critical
   role in investigating the localization and activity of nuclear R
   proteins and are useful for identifying the location of the nucleolus.
   However, they did not fully address potential impacts on protein
   localization in other cellular compartments or the functional influence
   of the fusion tag protein.

   Mass spectrometry remains essential for nucleolar proteomics research
   but is hindered by the challenges of isolating nucleolar proteins.
   Currently, the isolation of nucleolar proteins is exclusively
   accomplished via density gradient centrifugation ([62]23, [63]24),
   which is labor intensive and prone to sample degradation. A recent
   innovation, light-induced targeting of endogenous condensates combined
   with the proximity labeling technique BiolD, has identified five
   notably enriched endogenous proteins associated with RNA polymerase II
   ([64]25). The nucleolus, as the most essential endogenous condensate in
   eukaryotic cells, is characterized by transient constituent components
   and poses significant challenges for investigation using conventional
   biological assays such as mass spectrometry and other proteomics
   profiling techniques. Therefore, it is imperative to develop
   straightforward, effective, and precisely focused approaches for
   monitoring the interaction network and dynamics of nucleolar proteins,
   particularly in the context of cell stress and cancer.

   In this study, we developed an effective approach for real-time
   nucleolar localization in live cells and pioneered a proximity labeling
   technique targeting the nucleolar proteome. By recombinantly modifying
   the putative nucleolar targeting sequence of Legionella pneumophila
   effector protein Ceg10, we successfully engineered a specific signal
   peptide [termed NoB (nucleolar beacon)] transported by importin-β
   (IMPβ) to the nucleolus. This peptide serves as a nucleolar marker for
   real-time cell research. In addition, we introduced a versatile
   proximity labeling system by fusing the NoB with biotin-catalyzing
   enzymes (UltraID and TurboID), which allowed the construction of a
   nucleolar protein interaction network. In conclusion, this study
   presents a user-friendly and intuitive technique to specifically
   identify the nucleolar localization in living cells while introducing
   an effective method to analyze the nucleolar protein interaction
   network and dynamic changes of nucleolar components in eukaryotic
   cells. It also proposes a potential strategy for the delivery of
   precision medicine for diseases associated with nucleolar dysfunction.

RESULTS

NoB: A robust nucleolus indicator for living cells

   The successful colonization of L. pneumophila depends on its ability to
   regulate host cell physiological functions through T4SS (type IV
   secretion system) effector proteins, some of which can penetrate the
   host nucleus ([65]26, [66]27). Previous studies have shown that Ceg10,
   a T4SS effector, inhibits host-conserved antimicrobial defenses
   ([67]28, [68]29). However, the precise physiological roles and host
   pathways regulated by Ceg10 remain unknown. In this study, subcellular
   localization experiments demonstrated that ectopically expressed
   EGFP-Ceg10 fusion proteins localized to the nuclei of human embryonic
   kidney (HEK) 293T cells (fig. S1A). The nuclear import of proteins
   generally involves nuclear localization signals (NLSs), linear motifs
   that bind specific karyopherin-β family of nuclear transport receptors
   to facilitate translocation. Sequence analysis using three independent
   NLS prediction tools confirmed that Ceg10 harbors NLS within its 26– to
   64–amino acid region (fig. S1B). Notably, the amino acids 41 to 45 form
   a classical MP-NLS (monopartite NLS), characterized by clusters of
   positively charged residues such as arginine (R) and lysine (K)
   ([69]30, [70]31). Localization studies further revealed that the
   ectopically expressed truncated Ceg10[1–64] fused to the N terminus of
   EGFP was able to enter the nucleus and exhibited nucleolus
   localization. A similar pattern was observed with the truncated
   Ceg10[38–50] ([71]Fig. 1A). These findings indicate that the 38– to
   50–amino acid region functions as a nucleolar localization signal
   (NoLS). To enhance nucleolar localization efficiency, we constructed
   tandem repeats of the NoLS peptide. All exogenously expressed EGFP
   fusion proteins were observed to be partially or fully localized in the
   same compartment as NCL (nucleolin), and this observation was supported
   by fluorescence colocalization analysis ([72]Fig. 1B). Moreover, we
   noted that the intensity of green fluorescence within the nucleolus
   increased proportionally with the number of concatenated sequences
   ranging from one to five repeats. However, no further intensity
   increase was observed beyond five repeats. Constructs fused to
   Flag–hemagglutinin (HA) or Myc tags retained nucleolar localization,
   confirming that the signal was functional regardless of tag type (fig.
   S1C). The 5× NoLS sequence, termed NoB ([73]Fig. 1C), consistently
   localized to the nucleolus in diverse mammalian cell lines, including
   A549, NIH3T3, PC9, and HeLa cells ([74]Fig. 1D). This establishes NoB
   as a robust and reliable indicator for nucleolar localization in
   mammalian cells.

Fig. 1. Subcellular localization and nucleolar targeting of Ceg10 truncations
and NoB in mammalian cells.

   [75]Fig. 1.
   [76]Open in a new tab

   (A) Representative fluorescence images of 293T cells transfected for 24
   hours with EGFP, Ceg10[1–65]-EGFP, and Ceg10[38–50]-EGFP. (B)
   Subcellular localization and colocalization with NCL of one to six
   tandem recombinant NoLSs (Ceg10[38–50]) fused with EGFP at the C
   terminus after a 24-hour transfection in 293T cells. (C) Schematic
   illustration of the amino acid sequence and position of Ceg10[38–50] in
   the AlphaFold3-predicted structure of Ceg10, alongside the construction
   strategy for NoB. (D) Subcellular localization and colocalization of
   NoB-EGFP with NCL in NIH3T3, A549, PC9, and HeLa cells following a
   24-hour transfection. Green fluorescence indicates the position of EGFP
   or EGFP-fused proteins, nuclear DNA was visualized with Hoechst
   staining (blue), and NCL was visualized using Cy3-conjugated
   anti-rabbit antibodies and NCL-specific rabbit antibodies (red). Yellow
   lines in the merged panels represent regions analyzed for spatial
   relationships between NoLS/NoB and NCL using ImageJ. Scale bars, 10 μm.

KPNA1/KPNB1: Mediators of NoB nuclear translocation

   Classical NLSs typically bind IMPα proteins, which act as adaptors for
   IMPβ-mediated nuclear transport. To elucidate the nuclear import
   mechanism of NoB, we examined interactions between NoB and IMPα/β using
   coimmunoprecipitation (Co-IP) assays. The results demonstrated that NoB
   interacted with KPNA1, KPNA2, and KPNA6 of the IMPα family, as well as
   IMPβ/KPNB1 ([77]Fig. 2A). To further determine which nuclear transport
   proteins are critical for NoB transport, we reduced protein expression
   levels using small interfering RNA (siRNA) or short hairpin RNA (shRNA)
   and examined the distribution of NoB-EGFP through nuclear and
   cytoplasmic protein fractionation ([78]Fig. 2, B to E). The results
   showed that the knockdown of KPNA1 or KPNB1 significantly reduced the
   nuclear localization of EGFP-NoB, whereas the knockdown of KPNA2 or
   KPNA6 had minimal effect. Conversely, overexpression of KPNA1 and KPNB1
   enhanced the nuclear localization of EGFP-NoB ([79]Fig. 2F).
   Immunofluorescence analysis following KPNA1 or KPNB1 knockdown further
   corroborated these findings ([80]Fig. 2, G and H). Collectively, these
   results indicate that the nuclear transport of NoB-EGFP primarily
   depends on KPNA1 and KPNB1, highlighting their roles as key mediators
   for nucleolar targeting.

Fig. 2. Nuclear import of NoB is mediated by the IMPα/β.

   [81]Fig. 2.
   [82]Open in a new tab

   (A) Co-IP of NoB-EGFP and IMPα/β. (B to E) Protein expression levels of
   NoB-EGFP, KPNA1, KPNA2, KPNA6, and KPNB1 in nuclear and cytoplasmic
   fractions of 293T cells following siRNA- or shRNA-mediated knockdown.
   (F) Immunoblot (IB) analysis of NoB-EGFP, HA-KPNA1, and HA-KPNB1
   expression in nuclear and cytoplasmic fractions after overexpression in
   293T cells. LMNB2 (lamin-B2) and α-tubulin served as nuclear and
   cytoplasmic controls, respectively. (G and H) Immunofluorescence
   analysis of nucleolar localization and the fluorescence ratio of
   cytoplasm to nucleus analysis of NoB-EGFP following KPNA1 or KPNB1
   knockdown in 293T cells for 24 hours. Scale bars, 10 μm. Green
   fluorescence indicates NoB-EGFP localization, nuclei were visualized
   with 4′,6-diamidino-2-phenylindole (DAPI) staining (blue), and
   Cy3-conjugated immunoglobulin G anti-rabbit antibodies were used to
   display the in situ KPNA1 and KPNB1 (red). Each dot represents one cell
   (n ≥ 90) from three independent experiments, and the data were
   presented as the means ± SD. siKPNA1 and siKPNA6, small interfering
   RNAs targeting KPNA1 and KPNA6, respectively. shKPNA2 and shKPNB1,
   short hairpin RNAs targeting KPNA2 and KPNB1, respectively. siControl,
   small interfering RNAs control. Data correspond to means ± 95%
   confidence and were analyzed using one-way analysis of variance
   (ANOVA), including all comparisons with Tukey’s correction.
   ***P < 0.001.

293T cell expressing NoB-EGFP and HeLa cell expressing NoB-EGFP: Reliable
cell lines for investigation of nucleolar function

   Visualizing the localization, morphology, and dynamics of the nucleolus
   in living cells is crucial for studying the function of the nucleolus
   throughout the cell cycle. To accomplish this, we developed a
   lentivirus packaging expression plasmid to screen for cell lines
   consistently expressing the EGFP fusion NoB protein. Among the 30
   strains tested, the most promising 293T and HeLa colonies, designated
   as TCN (293T cell expressing NoB-EGFP) and HCN (HeLa cell expressing
   NoB-EGFP), respectively, showed successful nucleolar localization.
   Immunofluorescence analysis revealed that NoB-EGFP localized to regions
   of the nucleolus marked by NPM1 and NCL, corresponding to the granular
   component (GC) and dense fibrillar component (DFC) regions ([83]Fig.
   3A). In addition, Treacher Collins syndrome protein (TCOF1), a key
   regulator of RNA polymerase I that bridges ribosomal processing and
   modification enzymes, localizes to the fibrillar center (FC) region of
   the nucleolus ([84]32). The overlap of green fluorescence from NoB-EGFP
   with red fluorescence from TCOF1 further confirmed the FC localization
   of NoB-EGFP ([85]Fig. 3B). These observations indicate that NoB-EGFP is
   present in all three major nucleolar subregions.

Fig. 3. Stable expression and nucleolar localization of NoB do not disrupt
the physiological functions of TCN and HCN cells.

   [86]Fig. 3.
   [87]Open in a new tab

   (A and B) Fluorescence images showing NoB-EGFP localization in the
   nucleoli of TCN and HCN cells, colocalizing with NPM1 (red), NCL
   (purple), and TCOF1 (red), which mark GC, DFC, and FC of nucleoli,
   respectively. Green fluorescence indicates NoB-EGFP localization,
   nuclei were visualized with DAPI (blue), and NPM1, NCL, and TCOF1 were
   visualized using Cy3- or Cy5-conjugated secondary antibodies. Yellow
   lines in merged panels show regions analyzed for spatial relationships
   between NoB-EGFP and NPM1, NCL, or TCOF1 using ImageJ. Scale bars, 10
   μm. (C) The ratio of the fluorescence of NoB in the nucleolus of TCN
   and HCN cells to that in the entire cell. Each dot represents the
   picture result of one field of different horizons (n = 20) from three
   independent experiments. (D) NoB-EGFP expression levels in TCN and HCN
   cells were determined by immunoblot assay, with overexpressed EGFP in
   293T cells serving as the control. LMNB2 was used as an internal
   control. (E and F) Relative proliferation ratios over 24, 48, and 72
   hours, determined via MTT assay. h, hours. (G and H) Representative
   flow cytometry results from the cell death assays. The ratio of total
   cell death was calculated by flow cytometry from three independent
   experiments. (I and J) Cell cycle overlay diagram of 293T or HeLa, 293T
   or HeLa supplemented with colchicine (0.125 μg/ml), and TCN or HCN
   cells were determined by flow cytometry. Percentages of cells in
   G[0]-G[1], G[2]-M, and S phases were analyzed by the ModFit software.
   Data are presented as means ± SD, with 95% confidence intervals.
   Statistical analyses were performed using one-way ANOVA with Tukey’s
   correction. P > 0.05 (n.s.), *P < 0.05, and ***P < 0.001.

   The statistical results of fluorescence localization analysis indicated
   that the proportion of fluorescence in the nucleolus could reach ~95%
   ([88]Fig. 3C). Western blot analysis showed that the expression levels
   of NoB-EGFP in TCN and HCN cells were lower than those of transiently
   expressed EGFP in 293T cells ([89]Fig. 3D). A 72-hour cell growth assay
   demonstrated no significant difference in proliferation ratio between
   TCN and 293T cells or HCN and HeLa cells ([90]Fig. 3, E and F). Flow
   cytometric analysis indicated that NoB-EGFP expression did not induce
   cell death ([91]Fig. 3, G and H). Furthermore, cell cycle analysis
   confirmed that TCN and HCN cells exhibited normal cell cycle
   progression compared with wild-type cells ([92]Fig. 3, I and J).
   Therefore, we successfully established the TCN and HCN cell lines,
   using NoB with nucleolar localization functionality as a reliable
   indicator of nucleolar morphology and integrity in living cells.

Dynamic nucleolar changes observed in TCN and HCN cells

   To observe the alterations in the nucleolus indicated by NoB during
   mitosis in TCN and HCN cells, we monitored the fluorescence
   distribution every 6 hours after synchronizing the cells with
   colchicine. Different mitotic stages were successfully captured. The
   localization features of NoB-EGFP showed both similarities and
   differences between TCN and HCN at various stages of mitosis ([93]Fig.
   4A). In TCN cells at the G[2]-S phase, NoB-EGFP fluorescence exhibited
   a rounder shape and a higher number of nucleoli compared to the
   G[0]-G[1] phase, whereas this difference was less pronounced in HCN
   cells. As mitosis began, during chromatin condensation, the green
   fluorescence of NoB-EGFP decreased. In TCN cells, NoB-EGFP was
   diffusely distributed throughout the cell, whereas in HCN cells, it was
   predominantly localized in the cytoplasm. Notably, NoB-EGFP did not
   exhibit the same chromosomal distribution pattern as NCL.

Fig. 4. Dynamic changes of the nucleolus during mitosis and under stress
conditions.

   [94]Fig. 4.
   [95]Open in a new tab

   (A) Fluorescence images of TCN and HCN cells at various mitotic stages
   after colchicine synchronization treatment. (B) Quantitative results
   concerning the number of nucleoli in TCN and HCN cells at different
   time points after synchronous treatment, based on a minimum of 40 cells
   per sample. (C) Micrographs of TCN and HCN cells after a 30-min
   infection with Lp02, A. baumannii, and S. aureus (multiplicity of
   infection = 20), followed by an additional 6-hour incubation period. (D
   to F) Quantification of nucleolus number, nucleolar area, and the
   nucleolar to nuclear area ratio in infected TCN and HCN cells (n ≥ 80).
   (G) Micrographs of TCN and HCN cells after a 5-hour treatment with
   actinomycin D (AMD), CX-5461, or CX-3543 (10 μg/ml). (H to J)
   Quantification of nucleolar number, nucleolar area, and the nucleolar
   to nuclear area ratio in drug-treated TCN and HCN cells (n ≥ 80). Green
   fluorescence indicates the position of NoB-EGFP, nuclei were visualized
   with DAPI staining (blue), NCL was visualized using Cy3-conjugated
   anti-rabbit antibodies and NCL-specific rabbit antibodies (red), and BF
   represents the bright-field image. Data are presented as the
   means ± SD, with confidence intervals of 95%. Statistical analyses were
   performed using one-way ANOVA with Tukey’s correction for multiple
   comparisons. Nonsignificant differences are denoted as n.s. (P > 0.05),
   while statistical significance is indicated as follows: *P < 0.05, **P
   < 0.01, and ***P < 0.001. h, hours.

   At metaphase, when chromosomes were preparing to divide, NoB-EGFP
   fluorescence nearly disappeared in both TCN and HCN cells until
   anaphase, when chromosomes were fully separated. At this stage, NCL
   fluorescence remained localized around the chromosomes. By the end of
   mitosis, as the cells formed two intact nuclei, NoB-EGFP reappeared
   within the nucleoli, colocalizing with NCL protein fluorescence. These
   observations were further supported by live-cell imaging of HCN cells
   using a confocal microscope, which confirmed the findings from the
   time-point analyses (movie S1). Quantitative analysis of nucleolar
   numbers at different time points revealed substantial fluctuations in
   TCN cells, ranging from one to nine nucleoli per cell ([96]Fig. 4B).
   The number of nucleoli increased 6 hours after synchronization, peaked
   at 12 hours, decreased to the initial level at 18 hours, and began to
   rise again at 24 hours. This fluctuation pattern suggests that a single
   mitotic cycle in TCN cells takes ~18 to 24 hours, with the S phase
   occurring around 12 hours postsynchronization. In contrast, HCN cells
   exhibited a more consistent range of nucleolar numbers (1 to 4), with
   slight increases observed at 6 and 18 hours. Furthermore, we
   investigated nucleolar changes in TCN and HCN following infections with
   L. pneumophila (strain Lp02), Acinetobacter baumannii, and
   Staphylococcus aureus. Both TCN and HCN cells exhibited an increased
   number of nucleoli after L. pneumophila or S. aureus infection
   ([97]Fig. 4, C to F). However, the nucleolar size in HCN cells was
   notably smaller compared to that in TCN cells. In contrast, A.
   baumannii infection led to an increase in nucleolar numbers in HCN
   cells but did not affect nucleolar numbers in TCN cells. Similar
   results were also observed in 293T and HeLa cells (fig. S2, A and B).
   In addition, we evaluated the effects of well-characterized nucleolar
   stressors, including actinomycin D (AMD), CX-5461, and CX-3543, on TCN
   and HCN cells ([98]Fig. 4, G to J). The results were consistent with
   previous findings, such as nucleolar disruption and alterations in
   nucleolar morphology ([99]33–[100]35). Similar results were also
   observed in 293T and HeLa cells (fig. S2, C and D). Collectively, these
   findings demonstrate that TCN and HCN are reliable cell lines for
   monitoring dynamic nucleolar changes, establishing NoB as an
   exceptional tool for nucleolar analysis.

NoB-based proximity labeling: Approach for nucleolar proteomics in mammalian
cells

   The conventional sucrose density gradient centrifugation method for
   exploring interactions between nucleolar proteins is hindered by
   specificity and labor-intensive protocols. Addressing this limitation,
   we developed an effective approach that combines the nucleolar
   localization capability of NoB with proximity labeling technology. We
   constructed a NoB-fused biotin-catalyzing enzyme expression vector and
   transfected it into 293T cells. Fluorescence analysis confirmed that
   both UltraID-NoB-EGFP and TurboID-NoB-EGFP localized to the nucleolus
   ([101]Fig. 5A). This localization pattern was consistent across other
   cell types, including HeLa, PC9, A549, and NIH3T3 cells (fig. S3).

Fig. 5. Proximity labeling of the nucleolar proteome by NoB-mediated nuclear
import of UltraID.

   [102]Fig. 5.
   [103]Open in a new tab

   (A) Fluorescence images of UltraID-NoB-EGFP localized to the nucleolus
   in 293T cells. Green fluorescence indicates the position of
   UltraID-EGFP or UltraID-NoB-EGFP, nuclei were visualized with Hoechst
   stain (blue), and NCL was visualized using Cy3-conjugated anti-rabbit
   antibodies and NCL-specific rabbit antibodies (red). (B) Experimental
   schematic of the procedure for UltraID-NoB-EGFP proximity labeling
   nucleolar proteome. (C) Immunoblot analysis of streptavidin affinity
   pull-down of biotinylated proteins from 293T cells expressing
   UltraID-EGFP or UltraID-NoB-EGFP. (D to F) Volcano plot, Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway classification, and
   the top 20 Reactome enrichments of differentially notable proteins in
   the nucleolar proteome obtained through UltraID-NoB-EGFP proximity
   labeling. Sig., significance; GTP, guanosine 5′-triphosphate. (G to I)
   KPNA1 and KPNB1 mediate nuclear translocation of UltraID-NoB-EGFP. The
   protein expression levels of UltraID-NoB-EGFP, HA-KPNA1, and HA-KPNB1
   in nuclear and cytoplasmic fractions of 293T cells were determined by
   immunoblot assay following their overexpression (G) or
   siRNA/shRNA-mediated knockdown (H and I). LMNB2 and α-tubulin served as
   nuclear and cytoplasmic controls, respectively. h, hours.

   Among the constructs, we selected UltraID-NoB-EGFP for its ability to
   catalyze rapid biotinylation ([104]36), and we used it to investigate
   the nucleolar proteome ([105]Fig. 5B). Western blot analysis
   demonstrated that UltraID-NoB-EGFP effectively biotinylated surrounding
   proteins, which could be enriched using streptavidin-coupled agarose
   beads ([106]Fig. 5C). Compared to the control UltraID-EGFP, the
   molecular weight distribution of biotinylated proteins was notably
   different. Mass spectrometry analysis identified 252 significantly
   enriched proteins (P < 0.05) (data S1), including 62 up-regulated
   [log[2] fold change (log[2]FC) > 1 and P < 0.05] ([107]Fig. 5D).
   Notably, many up-regulated proteins included ribosomal subunits such as
   PRL23, PRL4, PRL21, PRS16, PRS3A, PRL19, etc. Other enriched proteins
   included NOP56 and NOP58 (localized in the FC) and NCL (localized in
   the DFC) (data S1). Twenty up-regulated proteins were entirely absent
   in the control group, suggesting potential nucleolus-specific roles.
   Examples include ENO1, JPT2, RBM10, TTC1, and TMED10. Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis
   showed that these proteins were involved in diverse signaling pathways,
   with “translation” being the most abundant category under genetic
   information processing ([108]Fig. 5E). Reactome pathway analysis
   further highlighted significant enrichment in RNA metabolism, protein
   translation, and ribosomal small subunit pathways ([109]Fig. 5F).

   Mass spectrometry also identified nuclear transport proteins KPNA1,
   KPNA3, IPO9, CSE1L, and KPNB1. Co-IP experiments revealed that NoB
   predominantly interacts with KPNA1 and KPNB1, consistent with prior
   findings ([110]Fig. 2 and fig. S4A). Overexpression of KPNA3, IPO9, and
   CSE1L did not enhance the nuclear localization of NoB-EGFP in 293T
   cells (fig. S4B). To determine whether the nuclear transport of UltraID
   depends on NoB, we adopted the same methods of overexpression and
   knockdown of KPNA1 and KPNB1. Further validation showed that the
   nuclear localization of UltraID-NoB depended on KPNA1 and KPNB1.
   Knockdown of these proteins reduced the nuclear levels of UltraID-NoB,
   while their overexpression increased nuclear accumulation ([111]Fig. 5,
   G to I). In conclusion, we developed an effective method for capturing
   the nucleolar proteome of mammalian cells by combining NoB with
   proximity labeling. This approach relies on the nuclear import
   trafficking mediated by KPNA1 and KPNB1, offering an effective tool for
   nucleolar proteomics and expanding the understanding of nucleolar
   functions.

DISCUSSION

   The nucleolus, a vital membrane-less structure within the eukaryotic
   nucleus, plays a crucial role in cell growth and the regulation of
   stress responses. However, because of its dynamic nature, there is
   still a lack of methods to efficiently visualize nucleolar localization
   in living cells and to study dynamic changes in protein interaction
   networks within the cell nucleolus. In this study, we explored the
   subcellular localization of the L. pneumophila effector Ceg10 and
   identified the short peptide Ceg10[38–50] capable of nucleolar
   localization as anticipated by NLS or NoLS prediction websites
   ([112]Fig. 1A and fig. S1, A and B). The central attribute of the
   sequence is the characteristic MP-NLS motif (RRSKKR), which is flanked
   by glutamate residues and features a lysine residue at the terminal end
   of the Ceg10[38–50] sequence, resulting in a peptide rich in polar
   amino acids. Upon cleavage of Ceg10[38–50] into two shorter peptides
   (QERRSKKR and ELHCK), there was a substantial reduction in nucleolar
   localization efficiency, especially for Ceg10[46–50], underscoring the
   necessity of the intact sequence for functionality (fig. S1D). After
   five consecutive repetitions of NoLS (NoB), the efficiency of nucleolar
   entry is maximized, allowing it to penetrate the nucleoli of various
   mammalian cells ([113]Fig. 1, B to D). The enrichment of lysine and
   arginine is a fundamental feature of NLSs. The P1 position of the
   classical MP-NLSs is invariably occupied by lysine [K(R/K)X(R/K)]
   ([114]37, [115]38). On the other hand, NoLSs are characterized by a
   higher proportion of these two amino acids and longer sequence length
   compared to NLSs ([116]39, [117]40). Numerous studies have demonstrated
   that NoLSs facilitate the accumulation of proteins in the nucleolus
   through a charge-dependent mechanism rather than a sequence-specific
   mechanism ([118]39, [119]41, [120]42). This view is supported by the
   analysis of the general nucleolar protein properties of Trypanosoma
   brucei, which identified basic amino acids as the key protein feature
   associated with nucleolar partition ([121]43). Recent research has
   revealed that nucleolar proteins contain D/E tracts, K blocks + E-rich
   regions, and proteins harboring high-scoring K blocks + E-rich regions
   are enriched in the FC/DFC, followed by GC ([122]44). The NoB displays
   the aforementioned characteristics, in addition to the presence of two
   E’s along with their relatively high content of the positively charged
   amino acids. The formation of K blocks and E-rich regions contributes
   to nucleolar localization. Furthermore, AlphaFold3 predicts that NoB
   cannot form a stable secondary structure. Previous studies have
   demonstrated that interactions mediated by disordered regions promote
   the specific localization of proteins to condensates ([123]38,
   [124]45). Therefore, the nucleolar localization of NoB is probably
   achieved through its high proportion of polar amino acids, K blocks,
   E-rich regions, and conformational flexibility.

   We further investigated the potential pathways for NoB nuclear entry.
   Classical NLSs typically contain four to eight positively charged
   residues that interact with the IMPα proteins, serving as adaptors for
   transport to the nucleus by IMPβ ([125]31, [126]38, [127]46, [128]47).
   In 293T cells, interaction studies between NoB and IMPα/IMPβ proteins
   revealed that NoB interacts with KPNA1, KPNA2, KPNA6, and KPNB1
   ([129]Fig. 2A). However, overexpression and knockdown of these
   transporters showed that KPNA1 and KPNB1 had the most notable impact on
   the nuclear-cytoplasmic separation assay, while the effects of KPNA2
   and KPNA6 were less pronounced ([130]Fig. 2, B to F).
   Immunofluorescence analysis further corroborated these findings, which
   indicate that the nuclear transport of NoB-EGFP depends on KPNA1 and
   KPNB1 ([131]Fig. 2, G and H). As the nuclear import–related proteins
   KPNA1, KPNA3, IPO9, and CSE1L were identified in the nucleolus mass
   spectrometry results, we also analyzed their interactions with NoB.
   These analyses indicated that NoB did not bind to KPNA3, IPO9, or CSE1L
   (fig. S4A). Compared to KPNA1 and KPNB1, overexpression of KPNA3, IPO9,
   and CSE1L did not significantly enhance the nuclear localization of NoB
   (fig. S4B). To eliminate the difference in nuclear transport mechanism
   caused by the difference in the length of the signal peptide and the
   protein carried by it, we monitored the nuclear-to-cytoplasmic protein
   ratio after knocking down or overexpressing KPNA1 and KPNB1 in cells
   overexpressing UltraID-NoB-EGFP. The results were consistent with those
   observed for NoB ([132]Fig. 5, G to I). Together, the translocation of
   NoLS into the nucleus is mainly mediated by KPNA1 and KPNB1. The Human
   Protein Atlas ([133]proteinatlas.org) shows that KPNA1 is detectable in
   all cell lines ([134]48). This characteristic is likely a key reason
   for NoB’s nucleolar localization across various mammalian cell types.

   Given the strong nucleolar localization properties of NoB, we
   established 293T and HeLa cell lines (referred to as TCN and HCN,
   respectively) stably expressing EGFP-NoB. These cell lines maintained
   normal physiological functions (e.g., growth, division, and apoptosis)
   and exhibited dynamic changes in nucleolar localization during the cell
   cycle ([135]Figs. 3 and [136]4, A and B, and movie S1). We observed
   that the number of nucleoli exhibited periodic fluctuations, with HCN
   cells showing smaller amplitude variations and shorter fluctuation
   periods compared to TCN cells. This characteristic aligns with the
   higher proliferative capacity of cancer cells. Exposure to
   well-characterized nucleolar stressors induced the expected nucleolar
   stress phenotypes ([137]Fig. 4, G to J), further validating the
   reliability of TCN and HCN cells for monitoring dynamic nucleolar
   changes. Another intriguing finding was the distinct changes in
   nucleolar number and morphology following infections with L.
   pneumophila, A. baumannii, and S. aureus ([138]Fig. 4, C to F). The
   observations underscore the critical role of NoB in detecting nucleolar
   stress responses induced by bacterial infections. These results also
   highlight the differences in stress responses between TCN and HCN cells
   to various pathogens, suggesting potential disparities in the
   regulation of nucleolar stress responses between normal and cancer
   cells. In future studies, we aim to further investigate the specific
   regulatory mechanisms underlying these phenomena. Compared to
   traditional nucleolar tracking methods, such as fluorescent
   protein–based reporters, rRNA probes, or nucleolar dyes, NoB exhibited
   high specificity and avoided nonspecific cytoplasmic targeting. Its
   versatility also enables researchers to establish additional nucleolar
   localization indicator cell lines. Therefore, these findings highlight
   the potential application of NoB as a precise indicator of nucleolar
   localization in living cells.

   Proteomic analysis of nucleolar proteins is essential for functional
   studies. In recent years, proximity labeling techniques have been
   widely applied across various aspects of proteomics research,
   facilitating the capture of transient or weak protein interactions
   ([139]36, [140]49–[141]51). In this study, we were surprised to find
   that NoB has the ability to transport various tags and proteins into
   the nucleolus, including biotin-catalyzed enzymes such as UltraID and
   TurboID ([142]Fig. 5A and fig. S3). By integrating NoB with UltraID, we
   successfully captured the nucleolar proteome, and the pathway
   enrichment analysis further validated its specificity to target
   nucleolar proteins ([143]Fig. 5, B to F). The notably different
   proteins encompassed a large number of ribosome proteins, such as
   PRL23, PRL4, PRL21, PRS16, PRS3A, PRL19, etc. In addition, key proteins
   localized to the FC region of the nucleolus, including NOP56 and NOP58,
   the DFC region, such as NCL, and the nucleolar periphery, such as
   DEAD-box helicase 21 (DDX21), were also identified. These findings
   demonstrate the effectiveness of this method for capturing the
   nucleolar proteome. We further evaluated the localization properties of
   UltraID-NoB-EGFP and TurboID-NoB-EGFP in other cell lines, confirming
   their versatility in mammalian cells (fig. S3). Collectively, this
   research highlights the potential of combining NoB with proximity
   labeling as a tool for analyzing the dynamic nucleolar proteome in
   mammalian cells.

   In this study, we identified and engineered a signal peptide sequence
   that exhibits nucleolar localization functionality and versatility in
   mammalian cells. Subsequently, we established nucleolar indicator cell
   lines for functional investigations in living cells. Furthermore, the
   combination of signal peptide and proximity labeling technology
   provided an effective approach to obtain the nucleolar proteome,
   successfully capturing the nucleolar-associated proteome ([144]Fig. 6).
   These techniques offer effective tools for investigating nucleolar
   functions in living cells and present a potential application method
   for drug delivery.

Fig. 6. Schematic diagram of NoB as a tool for monitoring nucleolar
morphology and proteomics.

   [145]Fig. 6.
   [146]Open in a new tab

   (Left) NoB-EGFP is transported into the nucleus through the combination
   of KPNA1 and KPNB1 and eventually reaches the nucleolus, where it can
   be used for nucleolar monitoring in living cells. (Right) The biotin
   catalytic enzyme UltraID bound to NoB-EGFP is transported into the
   nucleus via KPNA1 and KPNB1 and lastly reaches the nucleolus for the
   proximity labeling of the nucleolar proteome.

   In conclusion, we designed a signal peptide sequence, NoB, that
   exhibits strong nucleolar localization functionality and universal
   applicability in mammalian cells. Using NoB, we established nucleolar
   indicator cell lines for studying nucleolar functions in living cells.
   Furthermore, by combining NoB with proximity labeling technology, we
   successfully captured the nucleolar proteome ([147]Fig. 6). These
   techniques provide effective and versatile tools for investigating
   nucleolar functions in living cells and present a potential strategy
   for drug delivery applications.

MATERIALS AND METHODS

Cell culture

   HEK293T/293T, NIH3T3, HeLa, A549, and PC9 cell lines were cultured in
   Dulbecco’s modified Eagle’s medium (VivaCell, C3113-0500) supplemented
   with 10% fetal bovine serum (VivaCell, 04-001-1A), penicillin (50
   U/ml), and streptomycin (50 mg/ml; VivaCell, 03-031-5B). Cells were
   maintained at 37°C in a humidified incubator with 5% CO[2]. Mycoplasma
   contamination was checked using a universal detection kit (Beyotime
   Biotechnology, C0297S).

Plasmid constructions

   For ectopic expression in mammalian cells, full-length Ceg10,
   Ceg10[38–50], mutant variants, and tandem repeats (1 to
   6× Ceg10[38–50]) were cloned into the pcDNA3.1 vector with N- or
   C-terminal EGFP, Myc, or Flag-HA tags. Biotinylation enzymes (UltraID
   and TurboID) fused with NoB-EGFP were also inserted into the pcDNA3.1
   vector for expression in mammalian cells. For lentivirus packaging
   screening, NoB with an N-terminal EGFP tag was inserted into pCS-CG.
   The construction of these plasmids was accomplished using the
   homologous recombination method as per the ClonExpress Ultra One Step
   Cloning Kit protocol (Vazyme, C115-01). All plasmids and primers used
   are listed in tables S1 and S2.

Transfection and Co-IP

   Plasmids were transfected into cells at 80% confluence using Lipo8000
   (Beyotime, C0533) following the manufacturer’s protocol. After 24
   hours, the cells were lysed in Western blot and IP cell lysis buffer
   (Beyotime, P0013) at 4°C for 20 min. Lysates were centrifuged at 2500
   rpm for 10 min at 4°C, and the supernatant was incubated with Protein
   A + G Magnetic Beads (Beyotime, P2108) preconjugated with GFP-specific
   antibodies for 1 hour at room temperature. The beads were washed five
   times with precooled lysis buffer, and bound proteins were eluted,
   resolved via SDS–polyacrylamide gel electrophoresis (PAGE), and
   analyzed by immunoblotting using antibodies listed in table S3.

Western blotting

   Cells were lysed at 4°C for 30 min with the radioimmunoprecipitation
   assay (RIPA) buffer (Beyotime, P0013B). Lysates were centrifuged at
   12,000 rpm, and the supernatant was subjected to SDS-PAGE. Proteins
   were transferred to membranes and analyzed by immunoblotting using
   antibodies against GFP, HA, glyceraldehyde-3-phosphate dehydrogenase
   (GAPDH), α-tubulin, lamin-B2, KPNA1, KPNA2, KPNA3, KPNA6, and KPNB1
   (table S3).

Immunofluorescence

   After 24 hours of plasmid transfection, cells were washed with
   phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde for 30
   min at room temperature. Following three washes with PBS, the cells
   underwent permeabilization using PBS–Triton X-100 (0.5%, v/v) for 10
   min. The samples were incubated overnight at 4°C with antibodies (1:200
   dilution in 5% bovine serum albumin; refer to table S4 for antibody
   details) and subsequently with the fluorescence-labeled secondary
   antibodies for 60 min at room temperature. Last, staining of the cells
   was carried out using Hoechst 33342 (1:1000; Beyotime, C1025), followed
   by three washes with PBS and the addition of 200 μl of fluorescence
   decay-resistant sealing tablets. The fluorescence microscope (Leica,
   SP8) was used to analyze Hoechst and immunofluorescence signals, and
   the images were processed with the ImageJ software ([148]52).

Stable cell line establishment

   The NoB-EGFP sequence was cloned into the pCS-CG vector and
   cotransfected with packaging plasmids psPAX2 and pMD2.G into 293T cells
   at a 4:3:1 ratio. After 48 hours, the supernatant was collected and
   filtered through a 0.45-μm filter membrane. Subsequently, 293T and HeLa
   cells were infected with 4 ml of the viral supernatant, followed by the
   addition of Polybrene (1 μg/ml) to enhance infection. After incubation
   for 6 hours, fresh medium was added, and cells were cultured for an
   additional 48 hours. Monoclonal cell lines exhibiting fluorescence were
   sorted using a flow cytometer.

MTT assay

   The cytotoxicity of NoB-EGFP in TCN and HCN cell lines was evaluated
   using the MTT assay. The cells were trypsinized and seeded onto 96-well
   plates at a density of 5 × 10^3 cells per well. After an overnight
   incubation, the cells were designated as 0-hour samples. Subsequently,
   cells cultured for 24, 48, and 72 hours were collected, and 10 μl of
   MTT solution (5 mg/ml) was added to each well and incubated for 4
   hours. After removing the medium, the formazan formed in the cells was
   dissolved using 150 μl of dimethyl sulfoxide by shaking at 60 rpm for
   10 min at room temperature. The absorbance optical density values were
   then determined at a wavelength of 490 nm using a multimode reader
   (Molecular Devices, SpectraMax P1). The proliferation cell ratio was
   calculated using the following formula: optical density at 490 nm
   (OD[490]) of samples/OD[490] of wild-type HEK293T and HeLa at time
   point zero.

Flow cytometric analysis of cell death and cell cycle

   Cell death and cell cycle detection were carried out following the
   manufacturer’s protocol of the Apoptosis and Necrosis Assay Kit
   (Beyotime, C1056) and Cell Cycle Kit (Beyotime, C1052). After
   incubation in six-well plates for 24 and 48 hours, cells were
   trypsinized and centrifuged at 1000g for 5 min at 4°C, followed by
   three washes with PBS. Subsequently, the cells were resuspended in 200
   μl of 1× binding buffer containing 5 μl of Hoechst and 5 μl of
   propidium iodide (PI) and incubated for 20 min at room temperature. For
   cell cycle analysis, trypsinized cells were centrifuged at 1000g for 5
   min at 4°C, washed three times with PBS, and fixed overnight at 4°C in
   75% ethanol solution. The cells were then washed again thrice with PBS,
   gently resuspended in 195 μl of binding buffer, and stained with 5 μl
   of PI before being incubated at room temperature in the dark for 20
   min. Last, the samples were analyzed using flow cytometry (Beckman
   Coulter, Gallios), and the data were processed using the Kaluza and
   ModFit analysis software to obtain accurate results.

Gene knockdown

   To suppress the expression of KPNA1, KPNA2, KPNA3, KPNA6, KPNB1, and
   KPNB2, siRNAs targeting KPNA1 and KPNA6 were synthesized by General
   Biol (China). shRNAs were constructed in the pLV3 vector. HEK293T cells
   were cultured in six-well plates and cotransfected with NoB plasmids
   along with siRNA (100 nmol/ml) or PLV3-shRNA using Lipo8000 (Beyotime,
   C0533). Proteins were analyzed by SDS-PAGE and immunoblotting assays
   after 48 hours. The sequences for all siRNAs or shRNAs are listed in
   table S4.

Subcellular fractionation

   The cytoplasmic and nuclear fractions were isolated using the protocol
   provided by the Nuclear and Cytoplasmic Protein Extraction Kit
   (Beyotime, P0028).

Preparation of proximity labeling samples and streptavidin affinity pull-down

   The method of preparation of adjacent labeled samples is mainly
   referenced to Cho et al. ([149]50), with appropriate changes. Briefly,
   cells were cultured in 10-cm dishes. The following day, the cells were
   transfected with the pCDNA3.1-UltraID-NoB-EGFP plasmid. After 24 hours
   of protein expression, the medium was replaced, and 50 μM biotin was
   added. The cells were then incubated for an additional 18 hours,
   followed by three washes with PBS. The cells were lysed using 1 ml of
   RIPA lysis buffer, and the lysates were centrifuged at 12,000 rpm for 5
   min at 4°C to collect the supernatant. The collected supernatant was
   added to streptavidin agarose beads (Beyotime, P2159) and incubated at
   room temperature with rotation for 1 hour. Last, the beads were washed
   three times by centrifugation at 600g for 5 min at 4°C using
   phosphate-buffered saline containing Tween-20 (PBST).

Liquid chromatography–tandem mass spectrometry and data analysis

   Proteomic analysis was performed as previously described ([150]53).
   Briefly, samples collected through streptavidin affinity pull-down (20
   μl of streptavidin agarose beads) were washed twice with 1 ml of
   NH[4]HCO[3] (50 mM) and then resuspended in a solution containing 100
   μl of NH[4]HCO[3] (50 mM) and 10 μl of dithiothreitol (100 mM). The
   mixture was incubated at 56°C for 30 min. Subsequently, 10 μl of 300 mM
   iodoacetamide was added, and the samples were incubated at room
   temperature in the dark for 15 min. Trypsin (1:50, trypsin/lysate
   ratio) was then added, and digestion was carried out overnight at 37°C.
   On the next day, trypsin was added again, and digestion continued at
   37°C for an additional 5 hours. Digests were centrifuged using a 3-kDa
   filter to ensure only peptides passed through. The concentrations of
   peptides were determined by means of a modified Lowry Protein Assay Kit
   (Sangon Biotech Co.). Approximately 20 μg of peptides were desalted
   using Pierce C18 Spin Columns (Thermo Fisher Scientific) according to
   the manufacturer’s directions. The peptides were analyzed using a Q
   Exactive HF Orbitrap MS (data are available via ProteomeXchange with
   identifier PXD059125). Protein identification and label-free
   quantification were performed using the Proteome Discoverer software
   (version 2.2) with default settings. Proteins with an abundance greater
   than 55 were considered significant. A Student’s t test (P < 0.05) was
   applied to identify differentially abundant proteins. Log[2]
   transformation was applied to the protein abundance ratios, with only
   proteins exhibiting a log[2] value of >1 or <−1 being deemed
   significant.

   The functions of differentially abundant proteins were retrieved from
   the LegioList database [151]www.uniprot.org/database/DB-0054) and
   UniProt ([152]www.uniprot.org/). Functional category annotations and
   gene essentiality data were obtained from these databases. Pathway
   enrichment analyses were conducted using the KEGG and Reactome
   platforms via the free online data analysis tool OE Biotech Tools
   ([153]https://cloud.oebiotech.com/#/bio/tools).

Statistical analysis

   The data were presented as the means ± SD, derived from three
   independent experiments, and analyzed using GraphPad Prism version 6.01
   software (San Diego, CA, USA). The significance of differences between
   groups was assessed using one-way analysis of variance (ANOVA) with
   Tukey’s correction for multiple comparisons with GraphPad Prism
   (version 6.01; GraphPad Software, USA), with a threshold of P < 0.05
   considered to indicate statistical significance.

Acknowledgments