TTLL6-mediated Polyglutamylation of PurA Maintains Colonic Crypt Integrity

  1. Vanessa Pires
  2. Chloé Pangault
  3. Conception Paul
  4. Pascal Finetti
  5. François Bertucci
  6. Emilie Mamessier
  7. Célia Jardin
  8. Nuttanid Numnoi
  9. Ilaria Cicalini
  10. Damiana Pieragostino
  11. Laura Lebrun
  12. Carsten Janke
  13. Stéphane Audebert
  14. Jean-Paul Borg
  15. Krzysztof Rogowski
  16. Valérie Pinet
  17. Michael Hahne

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

The tubulin tyrosine ligase–like (TTLL) family comprises enzymes catalyzing posttranslational modifications of tubulin, including glutamylation and glycylation. We previously described a critical role for the monoglycylase TTLL3 in colon. Here, we identified TTLL6 as the predominant polyglutamylase in the colon, specifically expressed in epithelial cells of distal and transverse segments. TTLL6 expression decreases during CRC progression, correlating with poor patient prognosis. Deletion of Ttll6 in mice resulted in elongated colonic crypts, expansion of stem and transit-amplifying compartments, and increased numbers of differentiated epithelial cells. Moreover, Ttll6 -deficient mice showed an elevated susceptibility to chemically induced colon carcinogenesis. Notably, we identified the nucleic acid–binding protein PurA as a novel TTLL6 substrate with both proteins mutually required for nuclear localization. Consistently, both nuclear polyglutamylation and PurA were present in the bottom compartment of control colons, but reduced in Ttll6 -deficient colons. These findings reveal a TTLL6-PurA axis being critical in maintaining colonic homeostasis.

Article activity feed

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary

    In this manuscript, the authors investigate the role of the tubulin polyglutamylase TTLL6 in maintaining colonic epithelial homeostasis and its potential role in colorectal cancer (CRC). Using transcriptomic analyses, mouse genetics, histology, and proteomics, the authors report that TTLL6 is highly expressed in colonic epithelial cells and decreases during CRC progression. Constitutive and epithelial-specific deletion of Ttll6 in mice leads to elongated colonic crypts, expansion of proliferative and stem cell compartments, and increased susceptibility to chemically induced colitis-associated carcinogenesis. Mechanistically, the authors identify the nucleic acid-binding protein PurA as a potential non-tubulin substrate of TTLL6. They propose that TTLL6-mediated polyglutamylation of PurA regulates its nuclear localization, thereby contributing to epithelial homeostasis in the colon. Together, the study suggests a TTLL6-PurA axis that may restrain early colorectal tumorigenesis.

    Major comments

    1. Evidence that PurA is a physiologically relevant TTLL6 substrate remains incomplete. A central conclusion of the manuscript is that PurA is a substrate of TTLL6 whose polyglutamylation regulates nuclear localization. While the authors present several lines of evidence (PolyE immunoprecipitation, co-transfection experiments, and mutagenesis of the PurA C-terminal glutamate residues), the physiological relevance of this modification remains somewhat indirect. For example, polyglutamylation of endogenous PurA in colonic epithelial cells is inferred but not directly demonstrated. The PolyE antibody detects glutamate chains but does not identify the specific modified protein in tissue. Direct evidence that PurA is polyglutamylated in vivo (e.g., MS identification of the modification site on PurA or PurA immunoprecipitation followed by PolyE detection) would strengthen the mechanistic claim. At present, the data convincingly show that TTLL6 can glutamylate PurA in an overexpression system, but the endogenous modification remains less clearly demonstrated.
    2. Mechanistic link between PurA localization and the epithelial phenotype is not established. The authors propose that loss of TTLL6 disrupts PurA nuclear localization and thereby alters epithelial homeostasis. However, the manuscript does not establish a causal relationship between PurA localization and the observed crypt phenotypes. Specifically, it is not shown whether PurA loss phenocopies Ttll6 deficiency in the colon. No experiments test whether restoring nuclear PurA rescues the Ttll6 phenotype. Downstream transcriptional or signaling pathways regulated by PurA are not explored. Thus, while the TTLL6-PurA relationship is intriguing, the study remains largely correlative with respect to functional consequences.
    3. Interpretation of the tumorigenesis data should be tempered. The authors conclude that Ttll6 deficiency promotes colon carcinogenesis. However, the tumor data appear somewhat limited. Increased tumor numbers are reported only at an early time point (day 40) and are described as a trend toward significance. By day 70, tumor numbers and sizes appear comparable between groups. The increased incidence of vimentin-positive crypts is interesting but does not clearly establish increased tumor burden. Given these results, the conclusion that TTLL6 restrains tumorigenesis may be stronger than supported by the data. The authors may wish to frame this as enhanced early tumor development or altered tumor progression rather than increased tumorigenesis per se.
    4. Expansion of multiple epithelial cell populations requires clarification. The authors report that Ttll6-deficient colons exhibit expansion of stem/progenitor compartments as well as increased numbers of differentiated cells (e.g., goblet cells and enterocytes). While these findings are interesting, the biological interpretation is somewhat unclear. For example, expansion of stem/progenitor compartments typically accompanies reduced differentiation rather than increased differentiation. It is not clear whether the increased numbers of differentiated cells reflect overall crypt enlargement or altered lineage allocation. Quantification of cell-type proportions rather than absolute cell numbers would help clarify whether differentiation programs are altered.
    5. Nuclear polyglutamylation requires further clarification The authors report nuclear PolyE staining in colonic epithelial cells and propose that this reflects polyglutamylation of non-tubulin substrates such as PurA. However, it is not clear whether other nuclear proteins could account for this signal. The specificity of the nuclear PolyE signal should be better validated. Additional controls (e.g., peptide competition or validation with alternative approaches) would strengthen the interpretation.

    Minor comments

    1. The manuscript would benefit from clearer distinction between tubulin vs non-tubulin glutamylation throughout the text.
    2. Some conclusions in the Discussion appear slightly overstated relative to the data (e.g., the role of the TTLL6-PurA axis in tumor suppression).
    3. The description of the Ttll6 mouse models (constitutive vs conditional deletion) could be clarified earlier in the Results section.
    4. Quantification methods for histological analyses (crypt length, cell counts, marker-positive cells) should be described in greater detail in the Methods.
    5. It would be useful to include representative images of PurA localization in control vs Ttll6-deficient colon tissue in the main figures.
    6. Several minor typographical issues appear throughout the manuscript and should be corrected during revision.

    Significance

    General assessment

    This study investigates the role of the polyglutamylase TTLL6 in intestinal epithelial biology and colorectal cancer. The identification of a potential non-tubulin substrate (PurA) is conceptually interesting and expands the emerging view that tubulin-modifying enzymes can regulate additional cellular proteins. The study combines mouse genetics, histological analysis, transcriptomic datasets, and proteomics, which together provide a substantial dataset supporting a role for TTLL6 in regulating crypt architecture and epithelial proliferation. However, the mechanistic link between TTLL6 activity, PurA modification, and epithelial homeostasis remains incompletely resolved. The tumorigenesis data also suggest only modest effects on carcinogenesis.

    Advance relative to previous literature

    Previous studies have linked members of the TTLL family primarily to microtubule regulation and ciliary biology. This work extends these findings by suggesting a tissue-specific function of TTLL6 in the colon, and the existence of non-tubulin substrates regulating epithelial biology. If further validated, the identification of PurA polyglutamylation could represent an interesting conceptual advance.

    The manuscript will likely be of interest to researchers working in cytoskeletal post-translational modifications, intestinal epithelial biology, colorectal cancer biology

    My expertise lies in cytoskeletal regulation, epithelial biology, and intestinal tissue organization, which are directly relevant to the central themes of this manuscript.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary: In this study, the authors investigate novel functions of the tubulin typrosine ligase-like protein 6 (TTLL6), which covalently adds glutamate residues to the C-terminus of a given protein. The authors have already published previously on this topic. In the current study, the role of TTLL6 in colon function and pathologies was investigated. The study consists of two major parts. In the first part, a mouse model is used to show that TTLL6 is expressed at elevated levels and activity in epithelial cells of the colon. A database search indicated that TTL6 expression positively correlates with prognosis of patients with colorectal cancer. The authors generated a TTLL6 KO mouse and showed that induced tumor growth at 40 days was more positive for vimentin in the crypts of these mice, which should correlate with tumor aggressiveness. This difference was not observed anymore after 70 days. Morphological analyses of the crypts showed that in TTLL6 KO mice the crypts increased in length, a difference in proliferation markers, and a change of cell types in the crypts was observed.
    In the second part, the authors used a modification-specific antibody to immunoprecipitate (IP) proteins modified by TTLLs. To identify TTLL6-dependently modified target proteins, they compared these results with IPs from TTLL6 KO mice. A total of 43 proteins were identified this way. Because of their similarity to the tubuline tail sequence, two of the enriched proteins, PurA and PurB, were further analyzed. The authors provide evidence that PurA but not PurB is modified by TTLL6, which as a result changes its subcellular localization. While the first part of this work provides convincing novel insights into TTLL6's function with potential pathological relevance, the second part raises some concerns. I would therefore tend to rate the quality of the first part significantly higher than the second part.

    Major comments:

    1. When considering the results of the induced colorectal cancer test, the only significant difference between WT and KO was the moderately higher expression of Vimentin (figure 5E-F). Since this is the main evidence for a pathological relevance of TTLL6 in cancer, it is important to understand how the quantification of Vimentin in the complex tissue shown in figure 5E was done. A detailed description of how these images were analysed and perhaps a table with raw data would be essential to convince the reader of the conclusions. In the currently presented form, I find the analyses not too convincing.

    2. Figure 7A: It was somewhat surprising that two of the least significant (PurA is just below the cutoff) were used for further analyses. Although the authors explain that both proteins have strong sequence similarity to the know TTLL6 target, tubuline, the C-terminal, genomically encoded protein sequence of PurA and PurB already contain several glutamates. This raises the concern that the polyE antibody in the IP possibly detected the non-modified C-terminal tail of PurA and PurB and that both proteins may not be modified by TTLL6. Because of this and the lower significance than other candidates, the authors should consider focussing on other hits (OPTIONAL). Besides being much more significant, they lack an accumulation of glutamates in their C-terminus (at least the ones I looked at). Alternatively, the concern of having potentially IP-ed unmodified proteins should be addressed.

    3. Figure 8A: this figure compares PurA with a modified PurA that lacks the C-terminal EEE stretch. The authors conclude that the subcellular localization is different between both and that the nuclear localization of WT PURA must be due to modification by the co-expressed TTLL6. There are two major concerns with this conclusion: Firstly, the expression of PurA without TTLL6 co-expression is a missing essential control. This would show if PurA itself is already predominantly located in the nucleus regardless of potential modifications (PurA seems to have different nucleocytoplasmic localization in different cell types). Secondly, both depicted cells look very different. In PurA the nuclei are much smaller and the cytoplasm seems also much smaller than in the PurA DDD-expressing cells. Furthermore, IF staining without proper quantification of several cells seem less than ideal for such conclusions. In case, the authors want to convincingly validate this conclusion such a quantification with several cells would be required. OPTIONAL: an alternative approach would be a nucleo-cytoplasmic fractionation experiment followed by a western blot.

    Figure 8B: it seems that the contrast between the images of the upper and lower panel is very different. For this reason, I find it difficult to follow the conclusions. However, even when ignoring this aspect, I have great problems coming to the same conclusions as the authors.

    Minor comments:

    1. In figure 3A it would help if the legend describes what exactly "Control (+ or -)" means.

    2. In figure 3E-F, a label inside of the figure (what is the red bar, what the blue) would help the reader to faster grasp the subfigures.

    3. Figure 7C-D: these experiments are based on strong overexpression of TTLLs, which might result in unphysiological modifications of PurA. I would suggest to include a note of caution in the discussion that this is a possibility.

    4. In the discussion (page 9, last paragraph), it is stated: "Our findings suggest that the polyglutamylation of PurA is essential for maintaining colonic homeostasis". I do not understand this statement, as this study does not provide any evidence that modification of PurA does play a functional role in the colon (expression itself is not an evidence for function importance or even being "essential"). I recommend to remove this statement.

    5. Not all abbreviations are introduced properly (like CRC).

    Significance

    In general, this study addresses a very interesting aspect - i.e. the covalent addition of multiple glutamate residues to the C-terminus of a target gene by the enzyme TTLL6. The authors convincingly show that this protein regulates the morphology and composition of crypts in subregions of the colon. This is certainly a new and important finding that expands our knowledge about the functional breadth of this class of enzymes. If convincingly validated (see major concerns), also the pathological relevance of this enzyme for cancer progression would be of general interest. However, this statement has to be considered with a note of caution as this is not my area of expertise.

    The validation of novel targets of TTLL6 after IP is - at this stage of the manuscript - not very convincing to me. In particular the claim that PurA does play a functional role in the TTLL6-dependent regulation (of crypts) is not justified by the data. However, given that the list of other candidates contains several important gene regulators, this work might have the potential to open up to open up the field for new research directions.

    The reviewer's areas of expertise: cell biology, biochemistry, histology.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary:

    This study shows that loss of TTLL6 affects colonic epithelial homeostasis (crypt architecture and proliferation/differentiation markers) and proposes that TTLL6 contributes to a nuclear glutamylation signal, with PurA presented as a candidate non-tubulin substrate. The authors also connect TTLL6 mRNA levels to human CRC progression and outcome. Overall, the observations are potentially interesting, but the manuscript currently does not establish a clear mechanistic link between TTLL6 activity, PurA, and the in vivo phenotype.

    Major comments:

    1. The "TTLL6-PurA link" framing is currently too strong. The pull-down data indicate multiple candidate substrates, and the study does not show that PurA is the key functional mediator of the epithelial phenotype. As written, the manuscript reads as though PurA is the central downstream effector, which is not yet supported. Requested change: Either add substrate-specific functional evidence (additional KO/rescue-type experiments) or soften the language throughout (title/abstract/discussion) to reflect that PurA is one candidate among several.
    2. PurA glutamylation should be demonstrated directly by MS. PolyE/GT335 immunoblotting and enrichment in PolyE pull-downs are suggestive, but they do not conclusively establish glutamylation of PurA at the C-terminal end (antibody specificity and/or co-precipitating glutamylated proteins remain possible explanations). Essential experiment: MS/MS identification of glutamylated residue(s) on PurA, ideally with evidence that the modification is TTLL6-dependent (WT vs KO or epithelial-inducible KO).
    3. Regional TTLL6 expression vs phenotype needs to be reconciled. TTLL6 expression is reported to be highest in distal colon, yet the strongest crypt-length phenotype appears in transverse colon (as presented). Proximal colon data are not shown in the main text. Requested revision: Provide complete regional analyses (proximal/transverse/distal) with consistent quantification and statistics, and discuss explicitly why TTLL6 expression levels and phenotype do (or do not) align.
    4. Several internal inconsistencies and missing statistics.
      • Fig. 1A vs 1B: CEC enrichment appears ~80-fold in panel A and ~4-fold in the panel B; if these reflect the same enrichment workflow, this discrepancy needs a clear explanation (normalization, starting material, ....).
      • Fig. 2A: statistics are missing.
      • Fig. 5D: the effects appear borderline; the conclusions should match the statistical support/significance. Requested revision: Ensure complete statistical reporting in the manuscript (n, definition of replicates, test used, p-values/thresholds) and avoid interpretive language where differences are not significant.
    5. PurA Localization claims would benefit from stronger imaging and quantification. For nuclear localization/redistribution conclusions (main Fig. 8 and related supplement), confocal imaging with Z-stacks (and orthogonal views) would be more convincing than representative single-plane images. In addition, conditions with PurA-only expression need a clear baseline description and quantification. Requested additions: confocal Z-stacks + blinded quantification of nuclear/cytosolic localization across replicates; ideally support with subcellular fractionation and quantitative immunoblotting.
    6. Overexpression artifacts should be considered more carefully. If TTLL6 has been described as an elongase in prior work (Mahalingan, NSMB, 2020, DOI: 10.1038/s41594-020-0462-0) high-level overexpression may generate non-physiological modifications or localization patterns. Requested revision: Soften conclusions drawn from overexpression experiments and provide appropriate expression controls and/or supportive evidence in more physiological settings.
    7. Mouse tumor data should be interpreted more cautiously relative to the human correlations. The human datasets suggest a correlation between TTLL6 mRNA levels and clinical features/outcome (including recurrence-free survival), which is potentially interesting. In contrast, the mouse CAC results appear modest/borderline and, in places, are interpreted as stronger evidence than the data support. Requested change: Avoid strong claims about TTLL6 promoting or suppressing tumor growth unless supported by robust, clearly significant differences and comprehensive burden metrics.

    Minor comments:

    • Every figure should clearly state n (biological vs technical), statistical test, and multiple-comparison correction where applicable.
    • Where effects are segment-specific, the text should reflect that specificity and avoid global statements.
    • The Discussion would benefit from a clearer separation of what is directly shown versus what is proposed (especially near the end).
    • TTLL6 expression is largely presented at the transcript level; it would help to make this explicit throughout and avoid wording that implies protein-level validation where it is not shown.

    Significance

    The manuscript has the potential to be of interest because it points to a possible role for TTLL6 in non-tubulin, nuclear glutamylation in the intestinal epithelium, and it links TTLL6 expression to human CRC datasets. At present, however, the broader impact is limited by (i) insufficient direct evidence that PurA is glutamylated in vivo and (ii) the lack of a causal connection between PurA and the epithelial phenotype. In addition, while the human data show correlations between TTLL6 expression and clinical parameters/outcome, the mouse CAC phenotype appears comparatively modest/borderline and should be interpreted with appropriate caution. With stronger biochemical validation (MS), improved localization quantification, and more restrained framing (or additional functional data), the work could appeal to readers in intestinal epithelial biology, post-translational modification biology, and CRC research.

    Expertise: enzymology; post-translational modifications; microscopy; cancer mechanisms.

  5. Version published to 10.64898/2025.12.15.694360 on bioRxiv