ACS Infectious Diseases
Letter
reacted with sodium azide, followed by Jones oxidation to form
the corresponding azido carboxylic acid linker. After activation
by an N-hydroxysuccinimidyl (NHS) ester, the azido linker
was attached to 1 through amide bond formation. The
resulting VanN-azide (7) was purified by HPLC (Scheme S2).
Second, 1,5-pentadiol was brominated to form 1,5-dibromo-
pentane before being converted into a diazido derivative by
reacting with sodium azide. Using the heterogeneous
Staudinger reduction, 1,5-diazidopentane was then partially
reduced to form the linker 1-azido-5-aminopentane, which was
then reacted with 1 in the presence of HBTU to afford VanC-
azide (8) before HPLC purification (Scheme S3). Finally,
using CuAAC, 7 and 8 were linked together with commercial
Cy5.5-alkyne (9) or SulfoCy5.5-alkyne (10) to afford the
Van probes VanN-Cy5.5 (11), VanC-Cy5.5 (12), and VanC-
Sulfo-Cy5.5 (13) (Figure 1b and Scheme S3). (Note: In some
CuAAC reactions, the cycloaddition efficiency can be
significantly improved by supplementation with tris[(1-(2-
ethoxy-2-oxoethyl)-1H-1,2,3-triazol-4-yl)methyl]amine
(TEOTA) as the Cu(I) stabilizing ligand.)
We noticed that VanN derivatives 7 and 11 lost their
acylated vancosamine moiety to a certain extent and were
particularly notable when trifluoroacetic acid (TFA) was used
as an additive in HPLC. In contrast, during the preparation
and biological experiments, VanC derivatives 8, 12, and 13
seemed stable, as the amino group of the vancosamine moiety
was intact. In the hydrolysis reaction, the formation of the
positive oxocarbenium intermediate on vancosamine is a key
step and will facilitate the breakage of the glycosylic bond
between vancosamine and glucose. Since the amino group of
vancosamine in Van can be protonated by TFA to form
ammonium ions, this positively charged functional group can
prevent the formation of the positive oxocarbenium
intermediate and therefore resist hydrolysis. However, when
amidated, such as in VanN derivatives 7 and 11, the amino
group of vancosamine lost its protonation ability, causing the
glycosylic bond between amidated vancosamine and glucose to
control 5 and probe 6 did not affect bacterial growth during
our uptake experiments. Compared to controls without 5 or 6,
the bacterial morphology and quantity did not change when 5
or 6 was added. To confirm the iron dependency of our probe
6, we compared the labeling by 6 in S. aureus under different
levels of iron-limitation, ranging from restricted to abundant
iron conditions.
Clearly, labeling was profound under restricted conditions
(200 μM DP) and decreased as the amount of supplemented
free iron increased (Figure S5). The labeling was still notable
in LB with only residual iron (0 μM Fe). When extra iron was
supplemented (10−100 μM Fe), the labeling was abolished.
The correlation between labeling and iron-limitation showed a
strong iron dependency of labeling by SA-DORS-FL (6),
suggesting that our probe targeted S. aureus through the
siderophore-mediated iron uptake pathway, as shown
previously.20
Then, with SA and Van probes in hand, we examined their
targeting efficiency and strain selectivity. When using probe 6
in combination with three other Van probes under iron-
limiting conditions, both VSSA SA113 and the clinical isolate
of VRSA NR-49120 could still be efficiently labeled by 6, while
non-Staphylococcus bacteria, including E. faecalis, B. subtilis, and
E. coli, could not be labeled. Notably, the labeling efficiency of
VSSA was better than the labeling efficiency of VRSA, possibly
due to the thicker cell walls in VRSA (Figures 3a, S6a and
S7a). In contrast, three Van probes behaved differently in
bacterial labeling. VanN-Cy5.5 (11) only weakly labeled non-
Van-resistant Gram-positive bacteria, including VSSA, E.
faecalis, and B. subtilis, causing a 10- to 110-fold fluorescence
SulfoCy5.5 (13) strongly labeled three non-Van-resistant
Gram-positive bacteria. However, both Cy5.5- containing 11
and 12 also more or less labeled VRSA and VRE. Notably
when VRE was treated with 12, the labeling was very high
(Figure S7b). In contrast, 13 only limitedly labeled VRSA and
VRE (Figure 3b), despite the non-negligible VRE labeling. In
particular, when comparing the labeling between VSSA and
VRSA by our three probes, 13 exhibited the highest labeling
intensity when targeting VSSA (110-, 500-, and 1000-fold for
8, 12, and 13, respectively) (Figures 3b, S6b, and S7b) and the
greatest intensity ratio of VSSA/VRSA (5.7 vs 1.8 and 42 for
making probe 13 a better candidate for our detection protocol
development.
Using fluorescence microscopy, we showed that 6
specifically targeted S. aureus under iron-limiting conditions
by adding 200 μM 2,2′-dipyrdine (DP). Other bacteria, E.
faecalis, B. subtilis, and E. coli, were not targeted (Figure 2). DP
is frequently used in iron limitation.19 The attachment of the
triazole click linker, PEG linker, and fluorescein did not affect
the targeting ability and selectivity of 6. In addition, 2−10 μM
Similar to the flow cytometry results, probe 6 labeled VSSA
and VRSA exclusively in fluorescence microscopy (Figures 3c,
targeting specificity. Additionally, all three Van probes, 11, 9,
and 13, successfully labeled VSSA, E. faecalis, and B. subtilis.
Nevertheless, 13 still exhibited the greatest fluorescence
intensity of VSSA labeling in the same image settings (Figures
The observed labeling of some Van probes to the D-Ala-D-
Lac-dominant VRSA and VRE seemed confusing and contra-
dictory to previous knowledge. After examining the dose-
dependent labeling of three Van probes, we observed that a
high dose (5 μM) might cause cell lysis. In addition, by
measuring MICs (Table 1), we noticed that 11 and 12 killed
VSSA as expected, but unlike Van, they also killed VRSA and
even E. coli at concentrations as low as 4−8 μM. However, 13
behaved similarly to Van, efficiently killing VSSA but not
Figure 2. Demonstration of selective bacterial labeling by SA-DORS-
FL (6) using fluorescence microscopy. Bacteria were treated with
probe 6 (2.5 μM probe at 37 °C for 4 h) with DP (200 μM). Scale
bar: 5 μm (full) or 0.5 μm (enlarged).
C
ACS Infect. Dis. XXXX, XXX, XXX−XXX