In vivo modification of native carrier protein domains

Abstract

Carrier proteins are central to the biosynthesis of primary and secondary metabolites in all organisms. Here we describe metabolic labeling and manipulation of native acyl carrier proteins in both type I and II fatty acid synthases. By utilizing natural promiscuity in the CoA biosynthetic pathway in combination with synthetic pantetheine analogues, we demonstrate metabolic labeling of endogenous carrier proteins with reporter tags in Gram-positive and Gram-negative bacteria and in a human carcinoma cell line. The highly specific nature of the post-translational modification that was utilized for tagging allows for simple visualization of labeled carrier proteins, either by direct fluorescence imaging or after chemical conjugation to a fluorescent reporter. In addition, we demonstrate the utility of this approach for the isolation and enrichment of carrier proteins by affinity purification. Finally, we use these techniques to identify a carrier protein from an unsequenced organism, a finding that validates this proteomic approach to natural product biosynthetic enzyme discovery.

Figure 1

In vivo labeling strategy. Cells are grown in the presence of azido-pantetheine (1) or fluorescent pantetheine (2). After uptake, the native CoA biosynthetic enzymes convert the pantetheine analogues to CoA analogues which then are appended to endogneous CPs via the native PPTase enzyme. After cell lysis azido-modified CPs can be detected by reaction with alkyne probes (3–5).

Figure 2

Detection of in vivo modified carrier proteins. (A) Cultures were grown with or without compound 1 and reacted with RRX-alkyne 4 after lysis. Distinct CP labeling is seen in E. coli (1), B subtilis 168 (2), B. subtilis 6051 (3), and S. oneidensis (4) as compared to negative controls (B) Labeling of E. coli ACP by 1 with visualization by DMAC-alkyne 3 was effective in both LB (lane 1) and M9 minimal media (lane 2) as compared to negative controls (lanes 2,4).

Figure 3

Analysis of type I FAS ACP labeling in SKBR3 cells (A) In vivo modification of the human type I FAS. Lysate from cultures grown with compound 2 (+) show fluorescent modification of the FAS megasynthase as compared to negative controls (left). Blot with anti-FAS antibody confirms the presence and location of FAS on the gel (right). (B) FACS analysis of SKBR3 cells grown with compound 2 shows 99% of the SKBR3 cells grown with 2 took up the compound. (C) (panel 1) Compound 2 (black) incubated with PanK + ATP for 15 minutes results in addition of a phosphate on the 4′ hydroxyl and a shift to lower retention time (grey). (panel 2) After 120 minutes the reaction is complete, with all of 2 converted to the phospho-analogue. (panel 3) Lysate from SKBR3 cells grown with 2 contains mostly phospo-analogue. No further conversion to the CoA analogue in the SKBR3 cells could be detected (Supplementary Info).

Figure 4

Affinity purification of ACP. Lysate from cultures of S. oneidensis MR1 grown with compound 1 and then treated with biotin-alkyne 5 were incubated with streptavidin agarose allowing for purification of modified CP (crude lysate lane 1, desaltting column flow lane 2, recovered ACP band—between white arrows—10 kDa lane 3). Cultures grown with compound 1 but not reacted with biotin-alkyne 5 (lanes 4,5,6) or grown without compound 1 (lanes 7,8,9) showed no protein enrichment after incubation with streptavidin agarose. The strong band below ACP is a contaminate from the resin and was seen under all elution condtions.

Figure 5

Identification of a carrier protein from and unsequenced organism. (A) Metaboilc labeling of B. brevis 8246 by 1 gives a band around 10 kDa not present in the negative control. (B) Tryptic MS analysis of this band using a database of 200 known bacterial ACP sequences identified a peptide fragment matching an ACP sequence. (C) Degenerate and arbitrary PCR allowed for sequencing of the gene encoding this peptide and identification as the fatty acid ACP.

Figure 6

Alignment of Bacillus ACP sequences. Alignment of the new ACP sequence from B. brevis shows high homology with known Bacillus ACP sequences.