Whole Cell Luminescence-Based Screen for Inhibitors of the Bacterial Sec Machinery

Abstract

There is a pressing need for new antibiotics to combat rising resistance to those already in use. The bacterial general secretion (Sec) system has long been considered a good target for novel antimicrobials thanks to its irreplacable role in maintaining cell envelope integrity, yet the lack of a robust, high-throughput method to screen for Sec inhibition has so far hampered efforts to realize this potential. Here, we have adapted our recently developed in vitro assay for Sec activity─based on the split NanoLuc luciferase─to work at scale and in living cells. A simple counterscreen allows compounds that specifically target Sec to be distinguished from those with other effects on cellular function. As proof of principle, we have applied this assay to a library of 5000 compounds and identified a handful of moderately effective in vivo inhibitors of Sec. Although these hits are unlikely to be potent enough to use as a basis for drug development, they demonstrate the efficacy of the screen. We therefore anticipate that the methods presented here will be scalable to larger compound libraries, in the ultimate quest for Sec inhibitors with clinically relevant properties.

Figure 1

Design of the whole cell NanoBiT protein translocation assay. (A) The test substrate (pSpy-pep86) is expressed in E. coli and exported to the periplasm by the native Sec machinery. (B) For the assay, cell growth is stopped by cooling, then EDTA and lysozyme are added to disrupt the outer membrane and cell wall, respectively. This allows 11S to interact with periplasmic pep86, giving a luminescent signal proportional to the amount of secreted pSpy-pep86. (C) The counter assay is performed exactly like the assay, but with the addition of Triton X. This also disrupts the inner membrane, giving a luminescent signal proportional to the total amount of pSpy-pep86. (D) Maximum luminescence signal for cells with uninduced pSpy-pep86 plasmid (unind; background), mSpy-pep86 (i.e., lacking the signal sequence and thus not exported; mSpy; negative control), and pSpy-pep86 (pSpy; experiment). Data are normalized to the maximum signal for pSpy, and represent the mean and SEM from three biological replicates. (E) As in panel (D), but for the counter-assay. Time courses from which the data in panels (D, E) are derived are shown in Figure S1.

Figure 2

Assay validation with control compounds. Maximum luminescence (L_max) from the assay (solid shapes) and counter-assay (open shapes) as a function of concentration, for various compounds. In each case, data are normalized to the untreated sample (dashed horizontal line), and show the average ± SEM from three biological replicates. Lines are best fits to the Hill equation (see the Materials and Methods section) for the assay (solid lines) and counter-assay (dotted lines). (A) CJ-21058, added from a 10 mM stock in DMSO. The orange lines show the effect of the same amount of DMSO (see panel B). (B) DMSO. (C) CCCP. The gray triangles show the cell density (determined by absorbance at 600 nm) as a function of CCCP concentration, and represent the mean ± SEM from six biological replicates. Note that CCCP is added from a 4 mM stock in DMSO; this gives 1.25% DMSO at 50 μM CCCP – too low to affect the assay (see panel B). (D) Sodium azide (NaN₃). (E) Polymyxin B (pink diamonds), tetracycline (mustard circles) and kanamycin (mauve triangles).

Figure 3

Identification of four lead compounds from a library of 5000. (A) Schematic representation of the screening process. (B–E) Assay and counter-assay titrations of the four successful hit compounds, as in Figure 2. The chemical structures of the hits are shown below.

Figure 4

Uncoupling activity of the four hit compounds. DiOC₂(3) fluorescence (relative to untreated cells) of the four lead compounds at three concentrations. Data are the average ± SEM from three biological replicates. The equivalent signal for 5 μM CCCP is shown as an orange line (with dotted line for SEM).