TGM6 is a helminth secretory product that mimics TGF-β binding to TGFBR2 to antagonize signaling in fibroblasts

Abstract

TGM6 is a natural antagonist of mammalian TGF-β signaling produced by the murine helminth parasite Heligmosomoides polygyrus. It differs from the previously described agonist, TGM1 (TGF-β Mimic-1), in that it lacks domains 1/2 that bind TGFBR1. It nonetheless retains TGFBR2 binding through domain 3 and potently inhibits TGF-β signaling in fibroblasts and epithelial cells, but does not inhibit TGF-β signaling in T cells, consistent with divergent domains 4/5 and an altered co-receptor binding preference. The crystal structure of TGM6 bound to TGFBR2 reveals an interface remarkably similar to that of TGF-β with TGFBR2. Thus, TGM6 has adapted its structure to mimic TGF-β, while engaging a distinct co-receptor to direct antagonism to fibroblasts and epithelial cells. The co-expression of TGM6, along with immunosuppressive TGMs that activate the TGF-β pathway, may minimize fibrotic damage to the host as the parasite progresses through its life cycle from the intestinal lumen to submucosa and back again. The co-receptor-dependent targeting of TGFBR2 by the parasite provides a template for the development of therapies for targeting the cancer- and fibrosis-promoting activities of the TGF-βs in humans.

Fig. 1. TGM6 domain structure, similarity to TGM1, and signaling activity.

a Comparison of the domain structure and amino sequence identity of TGM6 relative to TGM1. Residue numbering is shown above the shaded boxes corresponding to each domain. b TGF-β1-, TGM1-, and TGM6-simulated luciferase reporter activity in NIH-3T3 fibroblasts. Data shown are the mean and standard deviation of triplicate measurements from one of two experiments with similar results. c Structure of TGM1-D3 (PDB 7SXB). The sidechains of Arg¹⁹⁸, Ile²³⁸, Tyr²⁵³, and Lys²⁵⁴ shown to be critical for binding are displayed in blue. d Sequence alignment of TGM1-D3 and TGM6-D3. Residues shown to be essential in TGM1-D3 for binding TGFBR2 and are conserved in TGM6 are shaded yellow; essential residue that is non-conserved in TGM6 is shaded cyan; all other residues that are conserved in TGM1 and TGM6 are shaded gray (except for conserved cysteines, which are shaded red). Residues highlighted by the red box are proposed to underlie the differential affinity of TGM6-D3 and TGM1-D3 for TGFBR2 (see “Results” section and Fig. 8). Note, the residues numbers of TGM1-D3 and TGM6-D3 differ by 160 due the presence of D12 only in TGM1. Source data of (b) provided as a Source Data file.

Fig. 2. TGM6 binds TGFBR2 through D3.

ITC thermograms (left) obtained upon injection of TGFBR2 into TGM6 (a) or TGM6-D3 (c). Thermograms are overlaid as two (a) or three (c) experiments shaded purple and blue and purple, blue, and cyan, respectively. Mean integrated heats and standard deviation (top) and accompanying fit to a 1:1 binding model are shown to the right of the thermograms with the residuals (bottom) as a function of the molar ratio. The difference between the K_D for TGM6 and TGM6-D3 for binding TGBR2 is not statistically significant (two-sided unpaired t-test p value = 0.18; assuming n = 2 for TGM6 and n = 3 for TGM6-D3 replicate experiment count). b ¹H–¹⁵N HSQC spectra of ¹⁵N TGM6-D3 alone (red) overlaid onto the spectrum of the same sample containing a 1.2-fold molar excess of unlabeled TGFBR2 (blue). Shown below is an expansion of the boxed region with all titration points labeled as the molar ratio of ¹⁵N TGM6-D3:TGFBR2. d, e ¹H–¹⁵N HSQC spectra of ¹⁵N TGFBR1 alone (red) overlaid onto the spectrum of the same sample containing a 1.2-fold molar excess of unlabeled TGM6-D45 (blue) or TGM6:TGFBR2 complex (blue) (d and e, respectively). Source data of (a–e) provided through Figshare [10.6084/m9.figshare.28179359].

Fig. 3. TGM6-D3, and TGM1-D3, compete with TGF-β for binding TGFBR2.

a Titration of TGFBR2 into mmTGF-β2-7M2R in the absence (purple) or presence (blue) of 6 μM TGM6-D3 as the lower-affinity binder. b Titration of TGFBR2 into TGM6-D3 in the absence (purple) or presence (blue) of 6 μM TGM1-D3 as the lower-affinity binder. Thermograms in (a, b) are presented as single experiment without (purple) and with (blue) competitor. Each panel includes the thermograms (top), mean integrated heats ± SD and fitted isotherms (middle), and fitting residuals (bottom) for the associated titrations. The data were globally fit using a simple competitive binding model. Source data of (a, b) provided through Figshare [10.6084/m9.figshare.28179359].

Fig. 4. TGM6 is a potent inhibitor of TGF-β and TGM1 signaling in fibroblasts, but not T cells.

Inhibition of SMAD2/3 CAGA reporter stimulated by TGF-β1 (orange symbols) or TGM1 (blue symbols) in NIH-3T3 (a) or MFB-F11 (b) fibroblasts by increasing concentrations of TGM6. Smooth black lines correspond to the fit of the data to a dose-dependent inhibition of TGF-β1 or TGM1 signaling by TGM6. c Inhibition of the TGF-β1 (orange symbols) or TGM1 (blue symbols) induction of the Foxp3 transcription factor in murine splenic CD4+ T cells by increasing concentrations of TGM6. Data shown in (a, b) are mean ± standard deviation of triplicate measurements from one of two experiments with similar results. Data shown in (c) is the mean and standard deviation of triplicate measurements from one experiment. Source data of (a–c) provided as a Source Data file.

Fig. 5. TGM6 requires attachment of D3 to D45 to inhibit and does not bind the TGM1 co-receptor CD44.

Inhibition of TGF-β reporter stimulated by TGF-β1 (orange symbols) or TGM1 (blue symbols) in NIH-3T3 (a) or MFB-F11 (b) fibroblasts by increasing concentrations of TGM6-D3. Data could not be reliably fit to a dose-dependent inhibition model. c Inhibition of TGF-β CAGA-GFP reporter in NIH-3T3 fibroblasts by TGM6, TGM6-D3, TGM6-D45, or TGM6-D3 plus TGM6-D45. Data shown in (a) and (b) are mean and standard deviation of triplicate measurements of one experiment. Data shown in (c) are the mean and standard deviation of triplicate measurements from one of three experiments with similar results. ITC thermograms (d) and fitted isotherm (top) and residuals (bottom) (e) for two separate titrations of mCD44 into TGM1-D45. f ITC thermograms for a single titration of mCD44 into TGM6-D45. Source data of (a–c) and (d–f) provided as a Source Data file associated with the article or through Figshare [10.6084/m9.figshare.28179359], respectively.

Fig. 6. Structure of the TGM6-D3:TGFBR2 complex and mimicry of mammalian TGF-β.

a Overall structure of the TGM6-D3:TGFBR2 complex determined by X-ray crystallography at a resolution of 1.4 Å. TGM6-D3 and TGFBR2 are shaded olive green and lavender, respectively. Loops that are not modeled due to weak density are indicated by dashed lines. Sidechains of key interfacial residues are shown, as are the intramolecular disulfide bonds. b Structure of the TGF-β3:TGFBR2 complex at a resolution of 2.15 Å (PDB 1KTZ). TGF-β3 and TGFBR2 are shaded burnt orange and lavender, respectively. Sidechains of key interfacial residues are shown, as are the intramolecular disulfide bonds. Structural data are available through the RCSB PDB under accession code 9E9G.

Fig. 7. Interface contacts of TGM6-D3 with TGFBR2 and mimicry of mammalian TGF-β.

Ionic interaction with TGFBR2 Asp⁵⁵ with Arg⁹⁵ of TGM6-D3 (a) and Arg³⁹⁴ of TGF-β3 (b). TGFBR2 Glu⁷⁸ also has an ionic interaction with TGM6-D3 Arg³⁸, but it does not interact with TGF-β. Interaction of TGFBR2 Leu⁵⁰ and Ile⁷⁶ with the hydrophobic pocket on TGM6-D3 formed by Tyr⁹³, Ile⁷⁸, and Tyr⁸⁰ (c) and TGF-β3 formed by Tyr³⁹⁰, Val³⁹², and Trp³³⁰ (d). Ionic interaction with TGFBR2 Glu¹⁴² and Asp¹⁴¹ with Tyr⁸⁰ and Arg⁸² of TGM6-D3, respectively (e) and TGFBR2 Glu¹⁴² with Arg³²⁵ of TGF-β3 (f). In all panels, TGFBR2 is shaded lavender and sidechains of key interaction residues are shown. In panels (a), (c), and (e), TGM6 is shaded green, and, in panels (b), (d), and (f), TGF-β3 is shaded burnt orange. Structural data are available through the RCSB PDB under accession code 9E9G.

Fig. 8. TGM6 residues responsible for its high affinity for TGFBR2.

ITC fitted binding isotherms (top) and residuals (bottom) for titration of TGM6-D3 KSGT (a) or TGM1-D3 QRRG (b) into TGFBR2. Fitted binding isotherms are overlaid as two experiments shaded purple and blue. Inhibition of TGF-β1 signaling in MFB-F11 fibroblasts as detected by the conversion of p-nitrophenylphosphate by secreted alkaline phosphatase (c, d) or by pSMAD2 western blotting (e) by either TGM6 (c, e) or TGM6 KSGT (d, e). In the MFB-F11 assay, the cells were stimulated with 0, 40, or 200 pM TGF-β1 (blue, yellow, and orange symbols, respectively). Activation of TGF-β1 signaling in MFB-F11 fibroblasts as detected by the conversion of p-nitrophenylphosphate by secreted alkaline phosphatase (f, g) or by pSMAD2 western blotting (h, i) by either TGM1 or TGM1 QRRG (f, h) or by TGM1-D123 or TGM1-D123 QRRG (g, i). Data shown in (c, d, f, g) are mean and standard deviation of triplicate measurements from one experiment. Blots shown in (e, h, i) are from one experiment with α-tubulin serving as a loading control. Source data of (a, b) and (c, i) provided through Figshare [10.6084/m9.figshare.28179359] or the Source Data file associated with the article, respectively.