Uptake of unnatural trehalose analogs as a reporter for Mycobacterium tuberculosis

Abstract

The detection of tuberculosis currently relies upon insensitive and unspecific techniques; newer diagnostics would ideally co-opt specific bacterial processes to provide real-time readouts. The trehalose mycolyltransesterase enzymes (antigens 85A, 85B and 85C (Ag85A, Ag85B, Ag85C)) serve as essential mediators of cell envelope function and biogenesis in Mycobacterium tuberculosis. Through the construction of a systematically varied sugar library, we show here that Ag85 enzymes have exceptionally broad substrate specificity. This allowed exogenously added synthetic probes to be specifically incorporated into M. tuberculosis growing in vitro and within macrophages. Even bulky substituents, such as a fluorescein-containing trehalose probe (FITC-trehalose), were incorporated by growing bacilli, thereby producing fluorescent bacteria; microscopy revealed selective labeling of poles and membrane. Addition of FITC-trehalose to M. tuberculosis-infected macrophages allowed selective, sensitive detection of M. tuberculosis within infected mammalian macrophages. These studies suggest that analogs of trehalose may prove useful as probes of function and for other imaging modalities.

Figure 1. Ag85 structure and enzyme activity

(a) Ag85A, Ag85B and Ag85C catalyze transesterification of trehalose 1, TDM and TMM. (b–c) The X-ray crystal structure of Ag85B suggests significant modification of trehalose substrates may be tolerated without altering binding contacts. Ag85B binds trehalose (yellow) (PDB ID: 1F0P). (c) The active site of Ag85B bound to trehalose. C-2 and C-4′ hydroxyls both point out of the active site. C-6 points into the hydrophobic tunnel (blue) that is thought to accommodate the long fatty acyl chains of the mycolic acids. Measurements to nearest residue are indicated in angstroms (PDB ID: 1F0P). Figure generated in PyMOL, Version 0.99rc6 (http://pymol.org/pymol).

Figure 2. Trehalose is incorporated into M. tuberculosis in vitro and in macrophages

( a) Incorporation of ¹⁴C-trehalose into M. tuberculosis culture over 24 h. (b) ¹⁴C-trehalose was added either to M. tuberculosis–infected (i, iii, v), or to uninfected (ii, iv, vi) macrophages for 24 h. Supernatant was removed (“extracellular”) and the macrophages were lysed and bacteria harvested by centrifugation to give a supernatant (“cytoplasm”) and pellet (“intracellular”) fraction. Radioactive trehalose is found in the cytoplasm yet accumulates in M. tuberculosis within infected macrophages. (c) ¹⁴C-trehalose is incorporated into TMM and TDM in M. tuberculosis. Arrowheads show the migration position of authentic samples of TMM and TDM. In i, the culture was labeled for 24 h with ¹⁴C-acetate to label fatty acids. In ii, the culture was labeled for 24 h with ¹⁴C-trehalose. In iii, culture was treated with ¹⁴C-acetate and INH to inhibit formation of TMM and TDM. ¹⁴C-trehalose labels spots that comigrate with TMM and TDM. iv shows the radio TLC from dual labeling with FITC-trehalose and ¹⁴C-acetate. v shows the fluorescence intensity in lane iv. vi shows a FITC-trehalose standard. (d) The glycolipids extracted from iv that comigrate with TDM and TMM contain mycolic acids; 1–8 correspond to individual spots from lane iv of c that were excised, saponified and run on a second TLC. The glycolipids in spots 6 and 7 contain lipids that comigrate with authentic mycolic acids and give the characteristic banding pattern of the three major mycolates. M. tuberculosis, Mtb; trehalose, Tre. Data reflect mean values ± s.d. for four experiments.

Figure 3. Trehalose library

Compound numbers match those in the green-amber-red screen (Table 1). Key modifications are highlighted in red. Syntheses and characterization of all compounds are described in the Supplementary Methods and Supplementary Results.

Figure 4. Labeling of individual bacilli in vitro and within infected macrophages

(a) H37Rv M. tuberculosis are labeled with FITC-trehalose and show signification accumulation of probe at the bacterial poles and membrane. (b) FITC-trehalose labeling of the RFP-expressing strain of H37Rv. (c) A merge of the FITC-trehalose and the RFP channels. (d) The mean fluorescence of poles and midsections of H37Rv M. tuberculosis, labeled with FITC-trehalose, are quantified and show statistically significant differential labeling. RFU, relative fluorescence units. (e–g) FITC-trehalose–labeled H37Rv-infected J774 macrophages. (e) FITC-trehalose–labeled H37Rv. In the DIC image f, the bacilli are indicated by arrows. The merged image (g) shows colocalization of FITC-trehalose and M. tuberculosis. (h–j) J774 BCG-RFP–infected macrophages. (h) Green labeling from FITC-trehalose. (i) Red labeling from RFP. (j) Merged images. (k–n) Triple labeling colocalization studies with RFP, FITC-trehalose and anti-LAMP antibody in H37Rv-RFP–infected BMMs activated with IFN-γ. Arrows indicate red bacteria that are not labeled with FITC-trehalose and that colocalize with LAMP-1. (k) The merged RFP and FITC images. (l) The labeling with anti-LAMP-1 (magenta) with colocalization indicated in white. (m) Three-dimensional reconstruction. (n) Quantitation of the colocalization between antibody signal and bacteria labeled with FITC-trehalose (green bars) or red bacteria (total population). P values for anti-LAMP-1: green versus green+IFN-γ (0.006), green versus total+IFN-γ (0.006), total versus green+IFN-γ (0.01), total versus total+ IFN-γ (0.001) and green+IFN-γ versus total+ IFN-γ (0.1). Error bars reflect s.e.m. for a minimum of three experiments. P values > 0.1 are not indicated. Scale bars, 5 μm. M. tuberculosis, Mtb; trehalose, Tre.