TTLL8 is a monoglycylase that initiates at multiple sites on α- and β-tubulin tails.

(A) Cartoon of the αβ-tubulin dimer and C-terminal tail sequences illustrating the chemical structure of monoglycine branches and polyglycine chains. In monoglycylation, the donor (incoming) glycine is linked to the γ-carboxyl group of an acceptor glutamate in the tubulin tail via an isopeptide bond (red). In polyglycylation, the donor glycine is linked to the α-carboxyl group of the terminal acceptor glycine in the growing glycine chain via a standard peptide bond (cyan). (B) Size exclusion column elution profile and Coomassie-stained SDS-PAGE gel of recombinant purified TTLL8 (C) Mass spectra of unmodified human microtubules incubated without (top) or with TTLL8 (bottom). The number of posttranslationally added glycines is indicated in green (for α-tubulin isoforms) and blue (β-tubulin isoforms). The weighted mean of the number of glycines (<nG>) added to α- and β-tubulin are denoted α + <nG>; β + <nG>. For experimental details see Materials and Methods. (D-E) Representative MS/MS spectra from two independent experiments for monoglycylated α- (D) and β-tubulin (E) C-terminal tail peptide fragments proteolytically released from microtubules glycylated by TTLL8. For additional spectra see Figure S2. (F) List of TTLL8 monoglycylation sites identified by MS/MS sequencing. The line marked with 2G indicates that a monoglycine can be present at any of the glutamates in that region.

TTLL10 is exclusively a tubulin polyglycylase that can only elongate existing branch-point glycines.

(A) Size exclusion column elution profile and Coomassie-stained SDS-PAGE gel of recombinant purified TTLL10. (B) Deconvoluted α- and β-tubulin mass spectra from unmodified human microtubules incubated with TTLL10 alone or with TTLL8 followed by incubation with TTLL10. The weighted mean of the number of glycines is denoted as in Figure 1. (C-D) Representative MS/MS spectra from two independent experiments for polyglycylated α- (C) and β-tubulin isoforms (D) C-terminal tail peptide fragments proteolytically released from microtubules glycylated by TTLL8 and TTLL10. For additional spectra see Figure S5. (E) List of TTLL10 polyglycylation sites identified by MS/MS sequencing.

TTLL10 requires monoglycylation for robust microtubule binding.

(A) Diagram of TIRF-based microtubule binding assays (B) Binding curve of TTLL10-SNAP(Alexa647) to unmodified microtubules, shown in magenta, and monoglycylated microtubules (α + 1.7G, β + 2.0G), shown in yellow with black outline. It was not possible to reach saturation in the reaction with unmodified microtubules. (C) TTLL10-SNAP(Alexa647) (cyan) association with unmodified microtubules (Hi-Lyte 488-labeled, magenta) and monoglycylated microtubules (unlabeled, yellow contour). The weighted mean of the number of glycines denoted as in Figure 1. For LC-MS of microtubules used in this assay see Figure S6. Assays performed at 500 nM TTLL10. Scale bar, 5 μm. (D) Quantification of TTLL10-SNAP(647) recruitment to differentially monoglycylated microtubules. Values normalized to unmodified microtubules; n = 66, 35, 39, and 51 microtubules from 14, 3, 5, and 6 independent experiments for the unmodified, β+1.2G, β+2.5G, and β+2.8G conditions, respectively. Statistical significance determined by Welsh’s t-test, with p < 0.0001 (****) or p < 0.001 (***).

TTLL10 recruitment to microtubules decreases with increased polyglycylation.

(A) TTLL10-SNAP(647) (cyan) association with monoglycylated microtubules (unlabeled, outlined in yellow) and polygylcylated microtubules (HiLyte 488-labeled, magenta). The weighted mean of the number of glycines denoted as in Figure 1. For LC-MS of microtubules used in this assay see Figure S7. Assays performed at 500 nM TTLL10. Scale bar, 5 μm. (B). Quantification of TTLL10-SNAP(Alexa647) recruitment of differentially polyglycylated microtubules normalized to levels on monoglycylated microtubules; n = 59, 20, 28, and 33 microtubules from 10, 3, 3, and 4 independent experiments for the β+2.8G, β+4.3G, β+5.4G, and β+7.7G conditions, respectively. Statistical significance determined by Welsh’s t-test, with p < 0.0001 (****) or p < 0.05 (*). (C) Representative images showing TTLL10-SNAP(647) association with monoglycylated and unmodified human microtubules as a function of time, in the presence of 1mM ATP with 1mM glycine (top two panels) and without 1mM glycine (bottom two panels). Scale bar, 5 μm. (D) Time courses of TTLL10-SNAP(Alexa647) recruitment to microtubules monoglycylated by TTLL8 (<nG>α ∼ 1.1, <nG>β ∼ 2.8) in the presence of 1 mM ATP with 1mM glycine (blue), 1 mM ATP without 1mM glycine (yellow), and unmodified microtubules in the presence of 1 mM ATP with 1mM glycine (black), and without 1mM glycine (green); n = 12 monoglycylated and 7 unmodified microtubules for the 1mM glycine time courses, and 8 monoglycylated and 8 unmodified microtubules for the glycine-free time courses. The same effect was observed across 3 independent experiments.

Polyglutamylation by TTLL6 enhances TTLL10 microtubule association.

(A) TTLL10-SNAP(Alexa647) (cyan) association with unmodified microtubules (HiLyte 488-labeled, magenta) and differentially modified microtubules (unlabeled, outlined in yellow): monoglycylated, monoglycylated/polyglutamylated, polyglycylated only, and polygylcylated/polyglutamylated. Monoglycylation added by TTLL8, polyglycylation by TTLL10 and polyglutamylation by TTLL6 (Materials and Methods). The weighted mean of the number of glycines denoted as in Figure 1. Glutamylation levels for glycylated/glutamylated microtubules are approximate and estimated by western blot (Figure S9) because quantification from LC-MS spectra of these dual-modified microtubules was not possible due to the large number of species present. Scale bar, 5 μm. (B) Quantification of TTLL10-SNAP(Alexa647) recruitment to modified microtubules normalized to those on unmodified microtubules; n = 84, 29, 30, 22, and 23 microtubules from 10, 4, 4, 3, and 3 independent experiments for monoglycylated, monoglycylated/polyglutamylated, polyglycylated, and polyglycylated/polyglutamylated microtubules, respectively. Statistical significance determined by Welsh’s t-test, with p < 0.0001 (****).

Extracted-ion chromatograms for monoglycylated α1b, βI, and βIVb tubulin tails modified by TTLL8.

(A-C). Extracted-ion chromatograms for monoglycylated α1b-tail peptides with one monoglycine branch (A), two monoglycine branches (B) and three monoglycine branches (C). (B-G). Extracted-ion chromatograms for monoglycylated βI-tail peptides with one monoglycine branch (D), two monoglycine branches (E), three monoglycine branches (F) and four monoglycine branches (G). (H). Extracted-ion chromatograms for monoglycylated βIVb-tail peptides. Tubulin tail peptides were proteolytically released from microtubules incubated with TTLL8.

MS/MS spectra for monoglycylated tubulin C-terminal tail peptides proteolytically excised from microtubules incubated with TTLL8.

(A, B). MS/MS spectra for monoglycylated α1b-peptides (A) and βI-tubulin C-terminal tail peptides (B).

Western blot analysis of glycylation in tubulin purified from tSA201.

0.2 μgs of purified tubulin were loaded for each unmodified and glycylated sample. Glycylation was detected with the Gly-pep1 and anti-polyG antibodies (Materials and Methods). We note a weak reactivity to the gly-pep1 antibody for the tubulin purified from tSA201 cells. It is unclear whether the antibody detects a very small proportion of mono- or bi-glycylated tubulin in this sample or the signal is due to a low affinity interaction of the antibody for the unmodified tubulin tail which is visible at these tubulin loading levels. Neither our LC-MS or MS/MS data of the tsA201 tubulin detected any glycylation on the intact tubulin, or tubulin tails, respectively. The higher molecular weight band is from TTLL8 which self-modifies during expression. The tSA201 tubulin shows no reactivity against the polyglycylation antibody.

Extracted-ion chromatograms for polyglycylated tubulin C-terminal tail peptides proteolytically excised from monoglycylated microtubules incubated with TTLL10.

(A-C). Extracted-ion chromatograms for polyglycylated α1b-peptides (A), βI-peptides (B) and βIVb-tubulin (C) C-terminal tail peptides. The subscript in Gi indicates the length of the polyglycine chain.

MS/MS spectra for polyglycylated tubulin C-terminal tail peptides proteolytically released from microtubules incubated with TTLL10 show elongation of polyglycine chains only at positions where monoglycine branches were already initiated by TTLL8.

(A, B). MS/MS spectra for α1b-(A) and βI-tubulin (B) C-terminal tail peptides. The subscript in Gi indicates the length of the polyglycine chain.

LC-MS of differentially monoglycylated microtubules used in assays shown in Figure 3.

The weighted mean of the number of glycines (<nG>) added to α- and β-tubulin are denoted α + <nG>; β + <nG>. The number of posttranslationally added glycines is indicated in green (for α-tubulin isoforms) and blue (β-tubulin isoforms) on the spectra.

LC-MS of monoglycylated and polyglycylated microtubules used in binding assays shown in Figure 4.

The weighted mean of the number of glycines (<nG>) added to α- and β-tubulin are denoted α + <nG>; β + <nG>.The number of posttranslationally added glycines is indicated in green (for α-tubulin isoforms) and blue (β-tubulin isoforms) on the spectra.

Western blot analysis of glutamylation in tubulin purified from tSA201 cells.

0.2 μg of tubulin was loaded for each unmodified and glutamylated sample. Glutamylation was detected with the GT335 antibody (Materials and Methods). No signal is detectable in the unmodified tsA201 purified tubulin. The MS/MS analysis of this tubulin did not identify any glutamylated peptides, either. In contrast, the tSA201 tubulin enzymatically glutamylated in vitro shows strong signal that increases with glutamylation level. Number of posttranslationally added glutamates indicated and determined from LC-MS measurements. The glutamylated tubulin runs slightly higher due to the changes in electrophoretic mobility.

Estimation of glutamylation levels of monoglycylated and polyglycylated microtubules used in microtubule binding assays shown in Figure 5.

(A) Western blot of polyglutamylated and monoglycylated or polyglycylated microtubules used in microtubule binding assays shown in Figure 5. From the left, first four lanes, TTLL6 polyglutamated microtubules with listed mean glutamate numbers determined from LC-MS spectra of intact microtubules. These were used to calibrate the polyglutamate signal detected with an anti-poly-E antibody (clone IN105, Adipogen; Materials and Methods); Subsequent four lanes, dually modified microtubules (glycylated with TTLL8 or TTLL8+10, and glutamylated with TTLL6) used in the assays shown in Figure 5; (B) Poly-E signal as a function of total glutamate numbers on α- and β-tubulin. The signal for the two dually modified species (monoglcylated + polyglutamylated and polyglycylated + polyglutamylated) used in the TIRF-based assays is shown with a discontinuous line.