Skip to main content
NCBI home page

This is a preprint.

It has not yet been peer reviewed by a journal.

The National Library of Medicine is running a pilot to include preprints that result from research funded by NIH in PMC and PubMed.

bioRxiv logoLink to bioRxiv
[Preprint]. 2026 Jul 14:2026.07.12.738044. [Version 1] doi: 10.64898/2026.07.12.738044

Live-cell co-translational folding tracking reveals bidirectional coupling between translation and folding

Rhiannon M Sears 1,†, Luis U Aguilera 1,†, Tristan M Bunting 1, Ning Zhao 1,*
PMCID: PMC13405136  PMID: 42523354

Abstract

Translation elongation and protein folding have long been proposed to coordinate during co-translational folding, yet the lack of technologies capable of simultaneously tracking both processes in live cells has hindered mechanistic understanding of this relationship. Here, we developed co-translational folding tracking (coTFT), a live-cell imaging platform that directly and simultaneously tracks translation and folding from individual mRNAs. Using reporters with distinct folding kinetics, we found that differences in folding kinetics were accompanied by corresponding changes in translation elongation rates. Conversely, altering translation elongation markedly affected protein folding outcomes. Combining coTFT with mathematical modeling enabled estimation of reporter folding times on translating ribosomes in live cells, confirming their distinct folding kinetics. Together, our results reveal that translation elongation and folding are bidirectionally coupled during co-translational folding.

Protein folding can begin while nascent chains are still being synthesized by ribosomes, a process known as co-translational folding (coF) (1–5). By coupling protein folding to the vectorial emergence of nascent chains from the ribosome, coF is thought to enhance protein folding efficiency (5–7). Translation elongation has long been proposed to regulate coF by controlling the timing of nascent chain emergence (8–12). Conversely, folding of nascent chains can generate mechanical pulling forces that feed back on the ribosome and potentially influence translation elongation (13). Despite decades of study, coF has been investigated primarily using in vitro systems (12, 13) or ensemble measurements (14, 15). Moreover, most existing assays infer folding from the function of mature proteins (16), providing indirect, endpoint readouts that cannot capture folding as it occurs on translating ribosomes. As a result, whether and how translation elongation and protein folding are coordinated during coF in live cells remains unresolved, limiting mechanistic insight into coF regulation.

To address this question, we developed co-translational folding tracking (coTFT), a live-cell imaging platform that enables direct and simultaneous tracking of translation and folding from individual mRNAs in their native intracellular environment. Applying coTFT to reporters with distinct folding kinetics reveals a novel regulatory principle in which translation elongation and protein folding are bidirectionally coupled to coordinate co-translational protein biogenesis in live cells.

Tracking single-mRNA co-translational folding in live cells

In this study, we developed a coTFT technology that enables direct visualization of protein folding on translating ribosomes from individual mRNAs in live cells. As illustrated in Fig. 1A, the coTFT reporter consists of an N-terminal spaghetti monster HA (smHA) tag containing 10 HA epitopes embedded within an mRuby scaffold (17), followed by tandem repeats of dark GFP (dGFP) (18). Upon translation initiation, a co-expressed anti-HA intrabody (ib) (19) fused to HaloTag-JF646 (anti-HA-ib-Halo) binds the smHA tag and labels nascent chains (Fig. 1A, magenta). As dGFP folds, a co-expressed conformation-sensitive (CS) anti-GFP intrabody fused to tandem mStayGold (mSG, anti-GFP-ib-mSG) selectively binds folded dGFP, marking nascent chains undergoing folding (Fig. 1A, green). CoF spots are identified as colocalized and co-moving nascent chain and folding spots.

Fig. 1. Tracking single-mRNA co-translational folding in live cells.

Fig. 1.

Open in a new tab

(A) Schematic of the co-translational folding tracking (coTFT) technology. The coTFT reporter consists of an N-terminal smHA tag fused to tandem dark GFPs (dGFPs). Nascent chain (NC) spots are labeled by a co-expressed HaloTag-fused anti-HA intrabody (anti-HA-ib-Halo, magenta), stained with JF646 dyes, which binds the 10 HA epitopes within smHA. Correctly folded NCs are visualized by a conformation-sensitive anti-GFP intrabody fused to tandem mStayGold (anti-GFP-ib-mSG, green). Co-translational folding (coF) spots are identified by colocalization of NC (magenta) and folding (green) spots. (B) Crystal structure of anti-GFP-ib (LaG16, tan) bound to GFPuv (gray) (PDB: 6LR7), highlighting the interaction interface (spheres) and illustrating recognition of a conformational epitope. (C) Left, representative FRAP images showing the photobleached region of interest (ROI; 3 μm diameter dark circle) in nuclei of cells expressing anti-GFP-ib-mSG (green) and 1×dGFP-mCh-H2B (magenta). Right, kymographs displaying fluorescence recovery over time along transverse lines (yellow, 50 pixels) across the bleached ROI shown in the left panels. Top, 1×dGFPFast; bottom, 1×dGFPSlow. (D) Normalized FRAP recovery curves of anti-GFP intrabody against dGFPFast and dGFPSlow within the ROI. Green, dGFPFast (mean ± SEM, n = 30 cells); black, dGFPSlow (mean ± SEM, n = 32 cells). (E) Box-and-whisker plots of the FRAP half recovery times. (F) Schematic of testing the designed linker for anti-GFP-ib-mSG binding to four tandem dGFPFast inserted between Mito and mCh (Mito-4×dGFPFast-mCh). The control construct Mito-mCh-mSG is shown at the top. (G) Representative images of cells expressing Mito-mCh-mSG (top) or co-expressing Mito-4×dGFPFast-mCh and anti-GFP-ib-mSG (bottom). (H) Box-and-whisker plots of normalized mitochondrial mSG to mCh fluorescence intensity ratios. From left to right: n = 74 cells (1.036 ± 0.027; mean ± SEM) and n = 60 cells (4.031 ± 0.287; mean ± SEM). (I) Schematic of the three coTFT reporters (6×dGFPFast, 6×dGFPSlow, and 6×dmCh) and corresponding western blot analysis confirming full-length protein expression. α-Tubulin served as a loading control. (J) Puromycin assay of the three coTFT reporters (mean ± SEM; n = 10 cells for each reporter). (K) Box-and-whisker plots of coF efficiencies for 6×dGFPFast (66.55 ± 2.68%, n = 17 cells, 2,334 trajectories), 6×dGFPSlow (25.84 ± 2.02%, n = 16 cells, 1,768 trajectories), and 6×dmCh (0.74 ± 0.24%, n = 20 cells, 2,987 trajectories) (mean ± SEM). In box-and-whisker plots, red lines indicate medians, **** p < 0.0001 (Mann-Whitney U test). Scale bars, 10 μm.

To determine whether coTFT can distinguish reporters with different folding kinetics, we selected fast- and slow-folding GFP variants: sfGFP and GFPuv (20–22) and quenched their fluorescence by introducing the G67A mutation (18, 23), generating dGFPFast and dGFPSlow reporters. We selected the high-affinity anti-GFP nanobody LaG16 (24) as the CS anti-GFP intrabody based on the crystal structure of the LaG16-GFPuv complex (PDB: 6LR7). The interaction interface spans multiple β-sheets on GFPuv (Fig. 1B), indicating that the epitope becomes accessible only after GFP adopts its folded conformation. Western blot analysis showed that the anti-GFP intrabody recognized the 6×dGFPFast reporter under native conditions but not after denaturation (fig. S1A), confirming its conformation-dependent binding and its ability to distinguish folding from non-folding nascent chains. Furthermore, the anti-GFP intrabody showed no detectable interaction with either smHA or mSG (figs. S1B–C), demonstrating its compatibility with the coTFT system.

To enable robust tracking of coF, the anti-GFP intrabody must bind both dGFPFast and dGFPSlow with high affinity. We therefore characterized its binding kinetics in live cells using fluorescence recovery after photobleaching (FRAP) (19, 25). In the FRAP assays, a defined nuclear region in cells co-expressing anti-GFP-ib-mSG and either 1×dGFPFast- or 1×dGFPSlow-tagged mCh-H2B was photobleached and fluorescence recovery was monitored over time (Fig. 1C, left). Because mCh does not interact with the anti-GFP intrabody (fig. S1B–C), it serves as an appropriate fluorophore for the FRAP assays. Given the extremely slow turnover of H2B (26), fluorescence recovery in the mSG channel primarily reflects intrabody exchange kinetics, with slower recovery indicating tighter binding. As expected, the magenta channels (1×dGFPFast- or 1×dGFPSlow-mCh-H2B) showed negligible recovery over ~10 min, confirming the slow turnover of H2B (Fig. 1C, right). In contrast, the green channel (anti-GFP-ib-mSG) displayed different recovery kinetics for the dGFP variants (Fig. 1D). The half-recovery times (t1/2) of the anti-GFP intrabody were 143.92 ± 9.08 s for dGFPFast and 75.89 ± 3.18 s for dGFPSlow (Fig. 1E), indicating slow turnover and therefore high-affinity binding to both variants. The slow turnover on dGFPSlow is consistent with the reported high affinity (KD = 6.7 nM) (24). The even slower recovery observed on dGFPFast suggests stronger binding, likely within the low-nanomolar or tighter range. Notably, the anti-GFP intrabody showed substantially slower recovery than the anti-HA intrabody (27), which is widely used for live-cell tracking of single-mRNA translation (19, 28–30). Together, these results indicate that the anti-GFP intrabody binds both dGFPFast and dGFPSlow with sufficient affinity to robustly label folding nascent chains as they emerge from the ribosome.

To increase fluorescence signals at coF sites, the reporter dGFP needs to be repeated multiple times using flexible linkers. We evaluated the linker design to determine whether it allowed maximal intrabody occupancy. In this assessment, cells co-expressing a mitochondrial outer membrane protein mitoNEET (Mito) (31) fused to 4×dGFPFast-mCh (Mito-4×dGFPFast-mCh) and anti-GFP-ib-mSG were imaged alongside control cells expressing Mito-mCh-mSG under identical conditions (Fig. 1F). Fluorescence intensities on mitochondria were quantified in both the mSG and mCh channels, and mSG/mCh ratio was normalized to the control (Fig. 1G). As shown in Fig. 1H, the normalized ratio increased from 1.036 ± 0.027 in control cells to 4.031 ± 0.287 in cells expressing Mito-4×dGFPFast-mCh, indicating that each Mito-4×dGFPFast-mCh molecule recruited approximately four anti-GFP intrabodies on average. These results demonstrate that the linker provides sufficient spacing to achieve near-maximal intrabody occupancy.

We next constructed three coTFT reporters: 6×dGFPFast, 6×dGFPSlow, and 6×dark mCh (6×dmCh), using the validated linker, generating transcripts of identical length (Fig. 1I, top). Because the anti-GFP intrabody does not interact with mCh (fig. S1B–C), the 6×dmCh reporter serves as an appropriate negative control. All three reporters produced full-length proteins after 24 h of expression (Fig. 1I, bottom). Proper folding of the reporter proteins was verified using the corresponding bright reporters (6×bGFPFast and 6×bGFPSlow), in which fluorescence was restored by introducing the A67G mutation into each dGFP repeat (figs. S1D–E). Both bright reporters were excluded from the nucleus, consistent with their large size and supporting expression of intact full-length proteins. Aggregation was minimal for both reporters, although aggregates were observed somewhat more frequently in cells expressing 6×bGFPSlow. Importantly, co-expressed anti-GFP-ib-Halo did not colocalize with these aggregates (fig. S1F), indicating that the anti-GFP intrabody does not recognize aggregated reporter proteins.

Using the established coTFT technology, we successfully visualized live-cell coF spots from individual mRNAs in cells expressing 6×dGFPFast and 6×dGFPSlow, but not 6×dmCh. These spots were confirmed as bona fide coF spots by their rapid disappearance following puromycin treatment, which releases nascent chains from ribosomes (Fig. 1J). To ensure accurate quantification, only puromycin-responsive coF spots were included in all subsequent analyses. We defined coF efficiency as the fraction of nascent chain spots that colocalized with folding spots. The 6×dGFPFast reporter showed a significantly higher coF efficiency (66.55 ± 2.68%) than 6×dGFPSlow (25.84 ± 2.02%), and both reporters displayed substantially higher coF efficiencies than the negative control 6×dmCh (0.74 ± 0.24%) (Fig. 1K). Together, these results establish coTFT as a platform for visualizing and quantifying coF on individual mRNAs in live cells and demonstrate its ability to distinguish proteins with distinct folding kinetics.

Fast- and slow-folding reporters exhibit corresponding differences in translation elongation rates

We next applied coTFT to address a long-standing question in coF: how translation elongation and protein folding are coordinated. To investigate this relationship, we employed dGFPFast and dGFPSlow, two GFP variants with identical lengths and ~97% protein sequence identity (fig. S2A) but distinct folding kinetics (20–22). To minimize confounding effects from codon usage and mRNA structure (32), we optimized their coding sequences to achieve ~98% nucleotide sequence identity (fig. S2B). Each dGFP variant was repeated four times using the validated linker and followed by two tandem dmCh repeats, providing additional time for the last dGFP domain to fold and be detected by the anti-GFP intrabody before ribosome release. An identical N-terminal smHA tag enabled nascent chain tracking and 24× MS2 stem-loops in the 3’ untranslated region (3’UTR) allowed mRNA tethering when needed. This design (4×dGFPFast and 4×dGFPSlow, Fig. 2A) controls for transcript length, codon usage, mRNA structure, translation ramping, and overall sequence composition (~99% nucleotide sequence identity and ~98% protein sequence identity), allowing us to specifically interrogate the relationship between translation elongation and protein folding.

Fig. 2. Fast- and slow-folding reporters exhibit corresponding differences in translation elongation rates.

Fig. 2.

Open in a new tab

(A) Schematic of the 4×dGFPFast and 4×dGFPSlow coTFT reporters, each consisting of an N-terminal smHA tag, four tandem dGFPFast or dGFPSlow repeats, and two tandem dmCh repeats. (B) Western blot analysis of the 4×dGFP reporters expressed for 24 h, confirming full-length protein expression. 4×FLAG-mCh-β-actin served as a transfection control and α-tubulin as a loading control. (C) Puromycin assay results for the two coTFT reporters (mean ± SEM; n = 12 cells for 4×dGFPFast and n = 15 cells for 4×dGFPSlow). (D) Box-and-whisker plots of co-translational folding (coF) efficiencies for 4×dGFPFast (71.62 ± 2.06%, n = 34 cells, 5,598 trajectories) and 4×dGFPSlow (28.00 ± 1.49%, n = 38 cells, 7,077 trajectories). (E) Box-and-whisker plots of the number of nascent chain (NC) trajectories per cell corresponding to the data in (D). (F) Box-and-whisker plots of the median NC intensity per cell corresponding to the data in (D). (G) Box-and-whisker plots of the median folding spot intensity per cell for 4×dGFPFast (4,259 trajectories) and 4×dGFPSlow (2,097 trajectories) corresponding to the data in (D). (H) Harringtonine run-off assay results for 4×dGFPFast (n = 26 cells) and 4×dGFPSlow (n = 28 cells). (I) Linear fit of the harringtonine run-off curve for 4×dGFPFast, resulting in a run-off time of 8.1 min and an elongation rate of 3.878 aa/s. (J) Linear fit of the harringtonine run-off curve for 4×dGFPSlow, resulting in a run-off time of 18.0 min and an elongation rate of 1.629 aa/s. In box-and-whisker plots, red lines indicate medians. * p < 0.05, **** p < 0.0001 (Mann-Whitney U test).

We confirmed by western blot that both reporters produced full-length proteins (Fig. 2B). Proper folding was verified using the corresponding bright reporters (4×bGFPFast and 4×bGFPSlow), which exhibited robust green fluorescence in live cells (figs. S3A–B). Similar to the 6×bGFP reporters, both bright 4×bGFP reporters were excluded from the nucleus, consistent with their large size and supporting the integrity of the full-length protein (figs. S3A–B).

For coF imaging, we generated a monoclonal stable cell line co-expressing anti-HA-ib-Halo and anti-GFP-ib-mSG under the control of a tetracycline-inducible promoter to minimize cell-to-cell variation in intrabody expression. Both reporters produced puromycin responsive coF spots (Fig. 2C). Consistent with the results obtained using 6×dGFP, the two reporters exhibited distinct coF efficiencies (movies S1 and S2), with 4×dGFPFast displaying a substantially higher coF efficiency (71.62 ± 2.06%) than 4×dGFPSlow (28.00 ± 1.49%) (Fig. 2D). The two reporters exhibited comparable numbers and intensities of nascent chains (Figs. 2E–F), whereas 4×dGFPFast displayed modestly brighter folding spots (Fig. 2G). This difference may arise from more folded dGFP domains on translating ribosomes in the 4×dGFPFast reporter, which was confirmed later in this study.

Harringtonine run-off assays revealed that the 4×dGFPFast reporter exhibited a more rapid decay than 4×dGFPSlow, indicating faster translation elongation (Fig. 2H). Linear fitting yielded elongation rates of 3.878 aa/s for 4×dGFPFast and 1.629 aa/s for 4×dGFPSlow (Figs. 2I–J). Because other major determinants of elongation were largely controlled in our reporter design and the presence of the anti-GFP intrabody did not affect elongation (figs. S3C–D), the observed elongation difference likely reflects the distinct folding kinetics of the reporters, suggesting that protein folding can influence translation elongation.

Perturbing translation elongation alters co-translational folding outcomes

Having established that protein folding can influence translation elongation during coF, we next investigated whether altering elongation affects folding outcomes. To perturb elongation, we employed the translation-pausing peptide XBP1u (S255A mutant, 26 amino acids) (33–36), which inhibits peptide-bond formation and thereby slows translation elongation. XBP1u was inserted either immediately upstream of the stop codon (4×dGFPFast-XBP1u) or between the N-terminal smHA tag and the first dGFPFast domain (XBP1u-4×dGFPFast) within the 4×dGFPFast reporter (Fig. 3A).

Fig. 3. Perturbing translation elongation alters co-translational folding outcomes.

Fig. 3.

Open in a new tab

(A) Schematic of the 4×dGFPFast reporters with or without the XBP1u translation pausing peptide. (B) Puromycin assay results for 4×dGFPFast-XBP1u (n = 12 cells) and XBP1u-4×dGFPFast (n = 13 cells). (C) Harringtonine run-off assay results for 4×dGFPFast (n = 26 cells, replotted from Figs. 2H and 2I), 4×dGFPFast-XBP1u (n = 22 cells), and XBP1u-4×dGFPFast (n = 24 cells). (D) Linear fit of the harringtonine run-off curve for 4×dGFPFast-XBP1u, resulting in a run-off time of 35.0 min and an elongation rate of 0.826 aa/s. (E) Linear fit of the harringtonine run-off curve for XBP1u-4×dGFPFast, resulting in a run-off time of 15.0 min and an elongation rate of 2.012 aa/s. (F) Box-and-whisker plots of co-translational folding (coF) efficiencies for 4×dGFPFast (71.38 ± 1.69%, n = 26 cells, 3,357 trajectories), 4×dGFPFast-XBP1u (80.28 ± 2.06%, n = 25 cells, 3,343 trajectories), and XBP1u-4×dGFPFast (15.25 ± 1.26%, n = 24 cells, 4,457 trajectories). (G) Box-and-whisker plots of the number of nascent chain (NC) trajectories per cell corresponding to the data in (F). (H) Box-and-whisker plots of the median NC intensity per cell corresponding to the data in (F). (I) Box-and-whisker plots of the median folding spot intensity per cell for 4×dGFPFast (2,475 trajectories), 4×dGFPFast-XBP1u (2,733 trajectories), and XBP1u-4×dGFPFast (675 trajectories) corresponding to the data in (F). (J-L) Models illustrating the expected ribosome density and spatial distribution on XBP1u-containing reporters relative to the parental 4×dGFPFast reporter. In box-and-whisker plots, red lines indicate medians, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 (Mann-Whitney U test).

We first confirmed that reporters containing XBP1u at either position produced full-length proteins, although at substantially lower levels than the reporter lacking XBP1u (fig. S4A). Puromycin treatment validated that the observed spots in XBP1u reporters represented coF events, as they rapidly disappeared upon puromycin treatment (Fig. 3B). To verify that XBP1u insertion did not impair GFP folding, we generated the corresponding bright reporters, 4×bGFPFast-XBP1u and XBP1u-4×bGFPFast, and successfully visualized green fluorescent cells expressing either reporter (figs. S4B–C). Notably, we observed puromycin-responsive nascent chain spots illuminated by folded and fluorescent GFP in cells expressing 4×bGFPFast-XBP1u. Given that bGFPFast requires ~10 min to fold and fluoresce, this observation suggests that XBP1u induces a substantial translational pause, allowing fluorescent GFP to accumulate while the nascent chain remains tethered to the mRNA.

We next measured the translation elongation rates of XBP1u-4×dGFPFast and 4×dGFPFast-XBP1u using harringtonine run-off assays. As shown in Fig. 3C, XBP1u-4×dGFPFast displayed a run-off time ~6.9 min longer than that of 4×dGFPFast (15.0 min versus 8.1 min), consistent with the ~6 min pause previously measured by socRNA (36). Interestingly, 4×dGFPFast-XBP1u showed an even longer run-off time than XBP1u-4×dGFPFast (35.0 min versus 15.0 min). The additional ~20 min delay may arise from the clearance of a ribosome traffic jam near the stop codon. Linear fitting of the run-off curves yielded elongation rates of 0.826 aa/s for 4×dGFPFast-XBP1u (Fig. 3D) and 2.012 aa/s for XBP1u-4×dGFPFast (Fig. 3E), both significantly slower than that of 4×dGFPFast (3.878 aa/s). These results demonstrate that both the presence and position of XBP1u strongly influence translation elongation.

Then, we quantified coF efficiencies for these reporters. As expected, 4×dGFPFast-XBP1u exhibited a significantly higher coF efficiency than 4×dGFPFast, increasing from 71.38 ± 1.69% to 80.28 ± 2.06% (Fig. 3F). This increase is most likely due to prolonged ribosome occupancy on the mRNA, which extends the observation window for coF events. In contrast, XBP1u-4×dGFPFast exhibited a markedly reduced coF efficiency (15.25 ± 1.26%) (Fig. 3F). Although similar numbers of nascent chains were detected for 4×dGFPFast-XBP1u compared to 4×dGFPFast and XBP1u-4×dGFPFast (Fig. 3G), both nascent chain and folding spot intensities were significantly higher for 4×dGFPFast-XBP1u (Figs. 3H–I). The elevated intensities are consistent with delayed ribosome release and increased ribosome occupancy relative to the other two reporters (Figs. 3J–K). In contrast, XBP1u-4×dGFPFast exhibited modestly more nascent chains than 4×dGFPFast (Fig. 3G), while nascent chain intensities remained comparable (Fig. 3H), suggesting similar ribosome loading (Figs. 3J, 3L). However, the substantially dimmer folding spots in XBP1u-4×dGFPFast (Fig. 3I) indicate a marked reduction in coF events per mRNA, likely due to a lower density of translating ribosomes downstream of the XBP1u sequence (Figs. 3J, 3L).

We applied the same XBP1u strategy to the 4×dGFPSlow reporter (fig. S4D). Consistent with the results obtained for the 4×dGFPFast reporters, XBP1u-containing reporters produced full-length proteins but at reduced abundance (fig. S4E). Quantification of puromycin-validated coF spots (fig. S4F) revealed that 4×dGFPSlow-XBP1u exhibited a markedly higher coF efficiency (70.04 ± 2.61%), whereas XBP1u-4×dGFPSlow displayed a substantially lower coF efficiency (7.47 ± 0.71%), than the parental 4×dGFPSlow reporter (29.67 ± 1.61%) (fig. S4G). Across all three reporters, nascent chain and folding spot intensities displayed trends comparable to those observed for the 4×dGFPFast reporters (figs. S4H–J).

Taken together, these results demonstrate that perturbing translation elongation substantially reshapes coF outcomes by altering the frequency of coF events on individual mRNAs. They further establish coTFT as a live-cell platform for real-time analysis of coF dynamics.

CoTFT reporters exhibit distinct co-translational folding kinetics in live cells

In developing coTFT, we selected dGFPFast and dGFPSlow based on their distinct in vitro refolding kinetics (20–22). However, their folding kinetics on translating ribosomes in live cells had not been quantitatively characterized. To address this knowledge gap, we generated a series of reporters containing increasing numbers (1× ~ 6×) of dGFPFast (Fig. 4A) or dGFPSlow (Fig. 4B) repeats and assessed how many dGFP domains could fold before translation termination.

Fig. 4. CoTFT reporters exhibit distinct co-translational folding kinetics in live cells.

Fig. 4.

Open in a new tab

(A-B) Schematic of the dGFPFast (A) and dGFPSlow (B) reporter series containing increasing numbers of dGFP repeats (1× - 6×). (C-D) Box-and-whisker plots showing co-translational folding (coF) efficiencies of the dGFPFast reporter series in (C) (1×dGFPFast: 31.29 ± 2.44%, n = 30 cells, 3,611 trajectories; 2×dGFPFast: 48.43 ± 3.06%, n = 32 cells, 3,702 trajectories; 3×dGFPFast: 66.60 ± 1.87%, n = 30 cells, 4,767 trajectories; 4×dGFPFast: 69.04 ± 1.52%, n = 32 cells, 3,777 trajectories; 5×dGFPFast: 69.65 ± 2.07%, n = 31 cells, 2,948 trajectories; 6×dGFPFast: 68.07 ± 1.91%, n = 34 cells, 4,331 trajectories) and the dGFPSlow reporter series in (D) (1×dGFPSlow: 15.36 ± 1.20%, n = 34 cells, 4,171 trajectories; 2×dGFPSlow: 21.02 ± 1.06%, n = 33 cells, 4,414 trajectories; 3×dGFPSlow: 26.98 ± 1.11%, n = 32 cells, 3,825 trajectories; 4×dGFPSlow: 24.86 ± 1.47%, n = 32 cells, 4,533 trajectories; 6×dGFPSlow: 24.61 ± 1.40%, n = 39 cells, 4,489 trajectories). (E-F) Fits of the folding time model to the experimentally measured coF efficiencies of the dGFPFast (E) and dGFPSlow (F) reporter series. (G) Predicted folding time of the 6×dGFPFast reporter, indicating that up to four dGFPFast repeats can fold before translation termination. (H) Predicted folding time of the 6×dGFPSlow reporter, indicating up to three dGFPSlow repeats can fold before translation termination. In box-and-whisker plots, red lines indicate medians. * p < 0.05, ** p < 0.01, **** p < 0.0001 (Mann-Whitney U test).

We first confirmed by western blot that all reporters produced full-length proteins (figs. S5A–B). We then quantified coF efficiency for each reporter (Figs. 4C–D). For both dGFPFast and dGFPSlow reporter series, coF efficiency increased as the number of dGFP repeats increased from 1× to 3×, followed by a plateau at higher repeat numbers (3× ~ 6×) (modestly decline between 3× and 6×dGFPSlow, p < 0.05). These results suggest that only the first three to four dGFP repeats have sufficient time to fold and be recognized by the anti-GFP intrabody before translation termination.

Fitting a mathematical model to the experimentally measured coF efficiencies (Figs. 4E–F) yielded estimated folding times of approximately 89 s for dGFPFast and 332 s for dGFPSlow (see Materials and Methods section), demonstrating that dGFPFast folds substantially faster than dGFPSlow on translating ribosomes in live cells. Using this model, we predict that the first four dGFPFast domains can fold before translation termination (Fig. 4G and table S2). However, the fourth folded dGFPFast domain is predicted to remain detectable for only a short time window (~27 s) before ribosome release, resulting in little contribution to the coF efficiency measurements and explaining the absence of a significant increase in coF efficiency between the 3× and 4×dGFPFast reporters. In contrast, the model predicts that only the first three dGFPSlow domains can fold before translation termination (Fig. 4H and table S2). This prediction is consistent with the dimmer folding spots observed for 4×dGFPSlow relative to 4×dGFPFast (Fig. 2G).

Together, these results demonstrate that coTFT can quantitatively resolve protein folding time during coF in live cells. The estimated folding times confirm that dGFPFast and dGFPSlow exhibit distinct folding kinetics on translating ribosomes in live cells, consistent with their previously reported relative folding behaviors in vitro (20–22).

Discussion

CoF couples protein synthesis with the acquisition of native protein structure, yet direct observation of folding on translating ribosomes in live cells has remained elusive. By developing coTFT, we provide an imaging platform that enables visualization and quantification of coF at the level of individual mRNAs in live cells. This capability allows mechanistic investigation of how translation elongation and folding are coordinated within their native cellular context.

Using coTFT, we found that fast- and slow-folding reporters, which differed by 3.7-fold in folding times, exhibited a corresponding 2.4-fold difference in translation elongation rates despite possessing nearly identical nucleotide and protein sequences and identical transcript lengths. The ~1.8% difference in protein sequence is unlikely to account for the observed elongation difference, as formation of the peptide bonds involving the substituted amino acids in dGFPFast and dGFPSlow (fig. S2A) is generally much faster than the overall elongation cycle and therefore is not rate-limiting (32, 37). Likewise, the nucleotide sequences differ by only 0.8% and contain highly optimized codons (CAI = 0.91, fig. S2B), indicating that codon usage is not a major contributor to the difference in translation elongation. Although subtle contributions from codon usage, mRNA structure, or sequence composition cannot be excluded, our findings suggest that protein folding can, in turn, influence elongation.

The mechanism underlying this feedback, however, is unlikely to involve folding-generated pulling forces near the ribosome exit tunnel (13), as kinetic analyses indicate that both reporters fold well beyond the tunnel exit (~345 aa for dGFPFast and ~542 aa for dGFPSlow). Instead, the distinct elongation rates may reflect differential interactions between folding nascent chains and the ribosome. Ribosomes are increasingly recognized as active participants in coF, interacting with nascent chains to stabilize or destabilize folding intermediates (12, 38–42). Such interactions may be more prolonged or frequent for the slow-folding reporter, thereby slowing elongation. Alternatively, the two reporters may differentially engage ribosome-associated chaperones, such as the nascent polypeptide-associated complex (NAC) (43, 44). By interacting with emerging nascent chains near the ribosome exit tunnel, NAC may modulate folding intermediates and thereby influence elongation. The molecular basis of this feedback remains to be elucidated, and the coTFT technology developed in this study should enable dissection of these mechanisms.

We further showed that altering translation elongation substantially changed coF outcomes, providing direct evidence that elongation influences folding. Notably, XBP1u-containing reporters produced markedly lower amounts of full-length protein despite exhibiting either increased or decreased coF efficiencies. Because prolonged XBP1u-mediated pausing can trigger ribosome collisions and activate ribosome-associated quality-control pathways (45, 46), the reduction in protein production is more likely attributable to collision-induced translational surveillance than to altered coF efficiencies.

Together, our work not only establishes coTFT as a novel technology for studying coF with single-mRNA resolution in live cells, but also provides direct evidence for a new regulatory principle in which translation elongation and protein folding are bidirectionally coupled to coordinate co-translational protein biogenesis in live cells.

Supplementary Material

Supplement 1
media-1.pdf (1.7MB, pdf)

Acknowledgments

We thank all members of the Zhao lab for valuable feedback on the manuscript. We also thank Drs. Timothy Stasevich, Judith Frydman, Hideki Taguchi, and Olivia Rissland for insightful discussion. This study was supported in part by the National Institutes of Health P30CA046934 funded Cancer Center Flow Cytometry Shared Resource [RRID:SCR_022035].

Funding

This work was supported by NIH grants K99GM141453 (NZ), R00GM141453 (NZ), R35GM160021(NZ), and Cystic Fibrosis Foundation grant 005749A123 (NZ).

Footnotes

Competing interests

Authors declare that they have no competing interests.

Data, code, and materials availability:

The complete code used for this study is available in GitHub: https://github.com/ningzhaoAnschutz/cof_paper.git.

Microscopy images available at: 10.5281/zenodo.20936426.

References and Notes

  • 1.Kleizen B., Van Vlijmen T., De Jonge H. R., Braakman I., Folding of CFTR Is Predominantly Cotranslational. Molecular Cell 20, 277–287 (2005). [DOI] [PubMed] [Google Scholar]
  • 2.Nilsson O. B., Hedman R., Marino J., Wickles S., Bischoff L., Johansson M., Müller-Lucks A., Trovato F., Puglisi J. D., O’Brien E. P., Beckmann R., von Heijne G., Cotranslational Protein Folding inside the Ribosome Exit Tunnel. Cell Reports 12, 1533–1540 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Frydman J., Erdjument-Bromage H., Tempst P., Hartl F. U., Co-translational domain folding as the structural basis for the rapid de novo folding of firefly luciferase. Nat Struct Mol Biol 6, 697–705 (1999). [DOI] [PubMed] [Google Scholar]
  • 4.Holtkamp W., Kokic G., Jäger M., Mittelstaet J., Komar A. A., Rodnina M. V., Cotranslational protein folding on the ribosome monitored in real time. Science 350, 1104–1107 (2015). [DOI] [PubMed] [Google Scholar]
  • 5.Kolb V. A., Makeyev E. V., Spirin A. S., Co-translational Folding of an Eukaryotic Multidomain Protein in a Prokaryotic Translation System*. Journal of Biological Chemistry 275, 16597–16601 (2000). [DOI] [PubMed] [Google Scholar]
  • 6.Ugrinov K. G., Clark P. L., Cotranslational folding increases GFP folding yield. Biophys J 98, 1312–1320 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Bitran A., Jacobs W. M., Zhai X., Shakhnovich E., Cotranslational folding allows misfolding-prone proteins to circumvent deep kinetic traps. Proc Natl Acad Sci U S A 117, 1485–1495 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Komar A. A., Samatova E., Rodnina M. V., Translation Rates and Protein Folding. Journal of Molecular Biology 436, 168384 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Collart M. A., Weiss B., Ribosome pausing, a dangerous necessity for co-translational events. Nucleic Acids Res 48, 1043–1055 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Ciryam P., Morimoto R. I., Vendruscolo M., Dobson C. M., O’Brien E. P., In vivo translation rates can substantially delay the cotranslational folding of the Escherichia coli cytosolic proteome. Proceedings of the National Academy of Sciences 110, E132–E140 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Oliver K. E., Rauscher R., Mijnders M., Wang W., Wolpert M. J., Maya J., Sabusap C. M., Kesterson R. A., Kirk K. L., Rab A., Braakman I., Hong J. S., Hartman J. L., Ignatova Z., Sorscher E. J., Slowing ribosome velocity restores folding and function of mutant CFTR. Journal of Clinical Investigation 129, 5236–5253 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Kim S. J., Yoon J. S., Shishido H., Yang Z., Rooney L. A., Barral J. M., Skach W. R., Translational tuning optimizes nascent protein folding in cells. Science 348, 444–448 (2015). [DOI] [PubMed] [Google Scholar]
  • 13.Goldman D. H., Kaiser C. M., Milin A., Righini M., Tinoco I., Bustamante C., Mechanical force releases nascent chain–mediated ribosome arrest in vitro and in vivo. Science 348, 457–460 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Nicola A. V., Chen W., Helenius A., Co-translational folding of an alphavirus capsid protein in the cytosol of living cells. Nat Cell Biol 1, 341–345 (1999). [DOI] [PubMed] [Google Scholar]
  • 15.Frydman J., Nimmesgern E., Ohtsuka K., Hartl F. U., Folding of nascent polypeptide chains in a high molecular mass assembly with molecular chaperones. Nature 370, 111–117 (1994). [DOI] [PubMed] [Google Scholar]
  • 16.Komar A. A., Unraveling co-translational protein folding: concepts and methods. Methods 137, 71–81 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Viswanathan S., Williams M. E., Bloss E. B., Stasevich T. J., Speer C. M., Nern A., Pfeiffer B. D., Hooks B. M., Li W.-P., English B. P., Tian T., Henry G. L., Macklin J. J., Patel R., Gerfen C. R., Zhuang X., Wang Y., Rubin G. M., Looger L. L., High-performance probes for light and electron microscopy. Nat Methods 12, 568–576 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Kutrowska B. W., Narczyk M., Buszko A., Bzowska A., Clark P. L., Folding and unfolding of a non-fluorescent mutant of green fluorescent protein. J Phys Condens Matter, 285223 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Zhao N., Kamijo K., Fox P. D., Oda H., Morisaki T., Sato Y., Kimura H., Stasevich T. J., A genetically encoded probe for imaging nascent and mature HA-tagged proteins in vivo. Nat Commun 10, 2947 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Pédelacq J.-D., Cabantous S., Tran T., Terwilliger T. C., Waldo G. S., Engineering and characterization of a superfolder green fluorescent protein. Nat Biotechnol 24, 79–88 (2006). [DOI] [PubMed] [Google Scholar]
  • 21.Crameri A., Whitehorn E. A., Tate E., Stemmer W. P., Improved green fluorescent protein by molecular evolution using DNA shuffling. Nat Biotechnol 14, 315–319 (1996). [DOI] [PubMed] [Google Scholar]
  • 22.Andrews B. T., Schoenfish A. R., Roy M., Waldo G., Jennings P. A., The rough energy landscape of superfolder GFP is linked to the chromophore. J Mol Biol 373, 476–490 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Reid B. G., Flynn G. C., Chromophore Formation in Green Fluorescent Protein. Biochemistry 36, 6786–6791 (1997). [DOI] [PubMed] [Google Scholar]
  • 24.Zhang Z., Wang Y., Ding Y., Hattori M., Structure-based engineering of anti-GFP nanobody tandems as ultra-high-affinity reagents for purification. Sci Rep 10, 6239 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Stasevich T. J., Mueller F., Michelman-Ribeiro A., Rosales T., Knutson J. R., McNally J. G., Cross-validating FRAP and FCS to quantify the impact of photobleaching on in vivo binding estimates. Biophys. J. 99, 3093–3101 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Kimura H., Cook P. R., Kinetics of core histones in living human cells: little exchange of H3 and H4 and some rapid exchange of H2B. J Cell Biol 153, 1341–1353 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Aguilera L. U., Chen S., Sears R. M., Yarbro J., DeRoo J., Ogg H. A., Geiss B. J., Stasevich T. J., Snow C. D., Zhao N., UTag, a cysteine-free thermostable tagging system for tracking single mRNA translation live. bioRxiv [Preprint] (2026). 10.64898/2026.05.06.723082. [DOI] [Google Scholar]
  • 28.Latallo M. J., Wang S., Dong D., Nelson B., Livingston N. M., Wu R., Zhao N., Stasevich T. J., Bassik M. C., Sun S., Wu B., Single-molecule imaging reveals distinct elongation and frameshifting dynamics between frames of expanded RNA repeats in C9ORF72-ALS/FTD. Nat Commun 14, 5581 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Cialek C. A., Galindo G., Morisaki T., Zhao N., Montgomery T. A., Stasevich T. J., Imaging translational control by Argonaute with single-molecule resolution in live cells. Nat Commun 13, 3345 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Galindo G., Fixen G. M., Heredia A., Morisaki T., Stasevich T. J., All Probes Plasmids (APPs) for multicolor and long-term tracking of single-mRNA translation dynamics. MBoC 36, mr6 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Colca J. R., McDonald W. G., Waldon D. J., Leone J. W., Lull J. M., Bannow C. A., Lund E. T., Mathews W. R., Identification of a novel mitochondrial protein (“mitoNEET”) cross-linked specifically by a thiazolidinedione photoprobe. Am J Physiol Endocrinol Metab 286, E252–260 (2004). [DOI] [PubMed] [Google Scholar]
  • 32.Rodnina M. V., Wintermeyer W., Protein Elongation, Co-translational Folding and Targeting. J Mol Biol 428, 2165–2185 (2016). [DOI] [PubMed] [Google Scholar]
  • 33.Yan X., Hoek T. A., Vale R. D., Tanenbaum M. E., Dynamics of Translation of Single mRNA Molecules In Vivo. Cell 165, 976–989 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Yanagitani K., Kimata Y., Kadokura H., Kohno K., Translational pausing ensures membrane targeting and cytoplasmic splicing of XBP1u mRNA. Science 331, 586–589 (2011). [DOI] [PubMed] [Google Scholar]
  • 35.Shanmuganathan V., Schiller N., Magoulopoulou A., Cheng J., Braunger K., Cymer F., Berninghausen O., Beatrix B., Kohno K., von Heijne G., Beckmann R., Structural and mutational analysis of the ribosome-arresting human XBP1u. eLife 8, e46267 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Madern M. F., Yang S., Witteveen O., Segeren H. A., Bauer M., Tanenbaum M. E., Long-term imaging of individual ribosomes reveals ribosome cooperativity in mRNA translation. Cell 188, 1896–1911.e24 (2025). [DOI] [PubMed] [Google Scholar]
  • 37.Rodnina M. V., Translation in Prokaryotes. Cold Spring Harb Perspect Biol 10, a032664 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Chan S. H. S., Włodarski T., Streit J. O., Cassaignau A. M. E., Woodburn L. F., Ahn M., Freiherr von Sass G. J., Waudby C. A., Budisa N., Cabrita L. D., Christodoulou J., The ribosome stabilizes partially folded intermediates of a nascent multi-domain protein. Nat. Chem. 14, 1165–1173 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Kaiser C. M., Goldman D. H., Chodera J. D., Tinoco I., Bustamante C., The Ribosome Modulates Nascent Protein Folding. Science 334, 1723–1727 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Burridge C., Waudby C. A., Włodarski T., Cassaignau A. M., Cabrita L. D., Christodoulou J., Nascent chain dynamics and ribosome interactions within folded ribosome–nascent chain complexes observed by NMR spectroscopy. Chemical Science 12, 13120–13126 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Liu K., Rehfus J. E., Mattson E., Kaiser C. M., The ribosome destabilizes native and nonnative structures in a nascent multidomain protein. Protein Sci 26, 1439–1451 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Samelson A. J., Jensen M. K., Soto R. A., Cate J. H. D., Marqusee S., Quantitative determination of ribosome nascent chain stability. Proceedings of the National Academy of Sciences 113, 13402–13407 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Lee J. H., Rabl L., Gamerdinger M., Goyal V., Khakzar K. M., Barbosa N. M., Abramovich J., Morales-Polanco F., Köhler A.-K., Samatova E., Rodnina M. V., Deuerling E., Frydman J., NAC controls nascent chain fate through tunnel sensing and chaperone action. Nature 652, 230–239 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Santos J., Günnigmann M., Gora R. J., Iljina M., Predin M., Kotan I. E., De P., Choudhary D., Jang J., Tippmann F., Hins C., Ban N., Tans S. J., Shan S., Kramer G., Bukau B., NAC promotes co-translational protein folding at the ribosomal tunnel exit. Molecular Cell 86, 1311–1326.e11 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Narita M., Denk T., Matsuo Y., Sugiyama T., Kikuguchi C., Ito S., Sato N., Suzuki T., Hashimoto S., Machová I., Tesina P., Beckmann R., Inada T., A distinct mammalian disome collision interface harbors K63-linked polyubiquitination of uS10 to trigger hRQT-mediated subunit dissociation. Nat Commun 13, 6411 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Han P., Shichino Y., Schneider-Poetsch T., Mito M., Hashimoto S., Udagawa T., Kohno K., Yoshida M., Mishima Y., Inada T., Iwasaki S., Genome-wide Survey of Ribosome Collision. Cell Rep 31, 107610 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Aguilera L. U., Raymond W. S., Sears R. M., Nowling N. L., Munsky B., Zhao N., MicroLive: an image processing toolkit for quantifying live-cell single-molecule microscopy. Bioinformatics Advances 6, vbag095 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Zhao N., Schmitt M. A., Fisk J. D., Phage display selection of tight specific binding variants from a hyperthermostable Sso7d scaffold protein library. The FEBS Journal 283, 1351–1367 (2016). [DOI] [PubMed] [Google Scholar]
  • 49.Lyon K., Aguilera L. U., Morisaki T., Munsky B., Stasevich T. J., Live-Cell Single RNA Imaging Reveals Bursts of Translational Frameshifting. Molecular Cell 75, 172–183.e9 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Kelkar D. A., Khushoo A., Yang Z., Skach W. R., Kinetic Analysis of Ribosome-bound Fluorescent Proteins Reveals an Early, Stable, Cotranslational Folding Intermediate. J Biol Chem 287, 2568–2578 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1
media-1.pdf (1.7MB, pdf)

Data Availability Statement

The complete code used for this study is available in GitHub: https://github.com/ningzhaoAnschutz/cof_paper.git.

Microscopy images available at: 10.5281/zenodo.20936426.

Articles from bioRxiv are provided here courtesy of Cold Spring Harbor Laboratory Preprints

RESOURCES