Volume 158, Issue 4, Pages 722-733 (August 2014) Phenotypic Variation of Salmonella in Host Tissues Delays Eradication by Antimicrobial Chemotherapy  Beatrice Claudi, Petra Spröte, Anna Chirkova, Nicolas Personnic, Janine Zankl, Nura Schürmann, Alexander Schmidt, Dirk Bumann  Cell  Volume 158, Issue 4, Pages 722-733 (August 2014) DOI: 10.1016/j.cell.2014.06.045 Copyright © 2014 Elsevier Inc. Terms and Conditions

Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 1 TIMERbac Fluorescence Properties (A) Schematic representation of TIMER fluorescence in nongrowing (upper panel) or actively growing (lower panel) cells. In nongrowing cells, both rapidly maturing green and slowly maturing orange TIMER molecules can accumulate. In dividing cells, rapidly maturing green molecules dominate over orange molecules that are diluted by cell division before maturation. (B) Agar plate with Salmonella constitutively expressing GFP, TIMERbac, or DsRed. (C) Fluorescence spectra of TIMERbac-Salmonella 18 hr after induction of TIMERbac expression (excitation spectra; left) or at various time intervals during maturation at 37°C, 5% O2, and pH 5.0 (emission spectra; right) (numbers show time points in hours). Similar data were obtained in two independent experiments. (D) Fluorescence maturation kinetics in TIMERbac-Salmonella at two different oxygen concentrations calculated from spectral data such as shown in (C). (E) Flow cytometry of Salmonella with constitutive expression of TIMERbac at defined division rates in chemostats maintained at 5% O2 (n.g., nongrowing culture). Similar data were obtained in three independent experiments. (F) TIMERbac-Salmonella fluorescence colors at different division rates and oxygen concentrations. Combined data from four experiments are shown (averages ± SD of two to four reactors for each data point). (G) TIMERbac-Salmonella fluorescence dynamics after switching division rates in chemostats (averages ± SD for two or three reactors from one experiment). Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 2 Salmonella Growth in Macrophage Cell Cultures (A) Fluorescence dilution analysis of GFP-loaded Salmonella 30 min (shaded areas) or 16 hr (red/blue lines) after infection of bone-marrow-derived macrophages with stationary (red; left panel) or exponentially grown (blue; right panel) in vitro cultures (under conditions suppressing SPI-1-mediated macrophage pyroptosis). Growth results in progressive GFP dilution and diminishing fluorescence intensities. Similar observations were made for the SPI-1 mutant SL1344 orgA and Maf-DKO macrophages in a total of eight independent experiments. (B) Fluorescence colors of TIMERbac-Salmonella 16 hr after infection of Maf-DKO macrophages with stationary (red) or exponentially grown (blue) in vitro cultures. Similar results were obtained for TIMERbac-SL1344 orgA as well as bone-marrow-derived macrophages in nine independent experiments. See also Movies S1, S2, S3, and S4. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 3 Fluorescence of TIMERbac-Salmonella in Infected Mouse Spleen (A) Flow cytometry of an infected mouse spleen homogenate (for gating, see Figure S1). Similar observations were made for more than 25 infected mice. (B) Fluorescence of wild-type (WT) TIMERbac-Salmonella (left panel) and TIMERbac-Salmonella ssrB deficient for expression of the SPI-2-associated type III secretion system. Green/orange fluorescence ratios for Salmonella in vitro cultures at 5% O2 and various division rates (Figure 1E) are shown in blue for comparison. Median green/orange ratios for replicates from individual mice are shown in Figure 6C. (C) Fluorescence colors of various attenuated TIMERbac-Salmonella strains in infected spleen at day 4 postinfection. Median green/orange ratios for replicates from individual mice are shown in Figure 6C. (D) Comparison of TIMER-based growth rate estimates with data from competitive infections (CIs, competitive indices). Data represent averages ± SD for one to five TIMER replicates and three to six CI values from individual mice (r, Spearman’s rank-order correlation coefficient). See also Figures S1 and S2. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 4 Salmonella Subsets with Distinct TIMERbac Colors in Infected Spleen (A) Flow cytometry sorting, reanalysis, and reinfection of TIMERbac-Salmonella subpopulations with different colors. Mice infected with sorted orange or green subsets were analyzed at day 4 p.i. (lower panel). This experiment was done once. (B) Confocal micrograph of an infected spleen cryosection. The insets show regions of interest with an overlay of F4/80 antibody staining (gray) recognizing resident red pulp macrophages. Additional micrographs show infected CD11bhi infiltrating cells (blue). The scale bars represent 5 μm. Similar observations were made in multiple sections from five independently infected mice. See also Figure S3. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 5 Extensive Variation of Salmonella Single-Cell Division Rates in Spleen (A) In vivo dilution of the nonreplicating plasmid pVE6007 (CAMR) in Salmonella subsets. Mice infected with TIMERbac-Salmonella/pVE6007 had a typical color distribution at 48 hr p.i. (upper panel; pooled data from four mice infected in two independent experiments). Salmonella subsets with different green/orange ratios were sorted and analyzed for the fraction of pVE6007-bearing CAMR clones (lower left panel; each data point represents a specific subset in one individual mouse; Exp, experiment). From these data, the number of divisions was calculated (lower right panel). Significance was tested with one-way repeated-measures ANOVA with posttest for linear trend. (B) In vivo division rate distribution calculated from in vivo colors and in vitro color-division rate relations at 5% O2 (averages ± SD of four independently infected mice). The blue line represents the FWHM, an expression of the extent of a distribution. Division rate estimates were compared to data from plasmid dilution for various Salmonella subsets (inset; averages ± SD of four mice). (C) Estimated daughter cell generation per day for Salmonella cells growing at various division rates (averages and SDs of four independently infected mice; based on data shown in B). Median and mean division rates are also shown. See also Figure S4. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 6 Properties of Salmonella Subsets (A) Proteome comparison of sorted green and orange TIMERbac-Salmonella subsets (see Table S1 for the full data set). Ribosomal proteins were more abundant in green subpopulations (∗∗∗p < 0.0001; two-tailed Wilcoxon signed-rank test), whereas σS-dependent proteins were more abundant in orange subpopulations (∗∗∗p < 0.0001). Many metabolic enzymes had differential abundance ratios. (B) Color distributions of parental strain SL1344 and various mutants in spleen. Similar data were obtained for another rcsC, another TA Δ3, two ssrB, four purH, and four SL1344 replicates. (C) Median green/orange ratios for various Salmonella strains (each symbol represents one mouse). (D) Abundance ratios of enzymes involved in biosynthesis of histidine (His) or purines (Pur), or degradation of Pur, sialic acid (Sial), or galactose (Gal) (∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; two-tailed t test on log-transformed data). See also Figure S5 and Table S1. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure 7 Impact of Antibiotic Therapy on Salmonella Subpopulations (A) Salmonella tissue loads during therapy with daily doses of 0.1 mg enrofloxacin. Spleen colonization levels before start of therapy (b; gray) were estimated based on flow cytometry counts. Each symbol represents one mouse. To avoid delays in TIMER flow cytometry, we prepared only the most distal small intestinal Peyer’s patch (Last PP) (mLN, mesenteric lymph nodes). Average disease scores ± SEM of between 20 (dose 1) and 3 (dose 5) mice from three independent experiments are also shown. (B) Division rate distributions at various time points during multidose therapy. Pooled data for two to five mice per time point are shown. (C) TIMER fluorescence 1 hr or 1 day after the first enrofloxacin dose. The red box indicates an antibiotic-induced new subset. Similar observations were made for 14 mice in five independent experiments. (D) Salmonella survival (CFUs per sorted TIMERbac-Salmonella cell) in subsets with different division rates from mice treated with one to three doses of enrofloxacin (averages ± SD for two to five mice from two independent experiments; one-way repeated-measures ANOVA with posttest for linear trend). (E) Proportions of live slow (s; <0.04 hr−1), moderate (m), and fast (f; >0.14 hr−1) Salmonella cells before or 1 hr after (“survivors”) daily enrofloxacin doses (averages ± SD of two to five mice per time point). At doses 4 and 5, low Salmonella loads prevented sorting-based analysis of survivor distributions. (F) Survival of Salmonella with various TIMERbac colors 1 hr after an oral dose of 1.3 mg ciprofloxacin in mesenteric lymph nodes in the streptomycin-pretreatment model (averages ± SD for three mice from two independent experiments; one-way repeated-measures ANOVA with posttest for linear trend). The inset shows the proportions of Salmonella with log(green/orange) <−0.055 among survivors in our typhoid fever/enrofloxacin model (En) and the streptomycin/ciprofloxacin model (CIP). (G) Fluorescence of Salmonella in mesenteric lymph nodes 1 hr or 1 day after an oral dose of 1.3 mg ciprofloxacin. (H) Color distributions for three mice from one experiment are shown. See also Figures S6 and S7. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S1 Flow Cytometry Analysis of TIMERbac-Salmonella, Related to Figure 3 All TIMER-expressing Salmonella contain green emitting molecules that are spectrally distinct from host tissue autofluorescence. We exploit this fact for gating. (A) Gating strategy for Salmonella with high TIMER content (pBR322 replicon). A tissue homogenate from mice infected with Salmonella SL1344 not expressing any fluorescent protein (right panel) confirms absence of any host background particle with similar fluorescence thus enabling to detect even few Salmonella in entire organ homogenates (Ex, excitation; Em, emission). (B) Gating strategy for Salmonella with moderate TIMER content (pSC101 replicon). Due to considerably overlap in the green channels, we use in addition the spectral properties of orange TIMER molecules to further separate TIMER signals from host autofluorescence. (C) Comparison of green fluorescence intensities for various strains. The single-copy chromosomal construct SL1344 virKp::timerbac emitted detectable fluorescence under in vitro inducing conditions, but the intensities were too low for in vivo detection. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S2 TIMERbac-Salmonella Colors in Different Infection Models, Related to Figure 3 (A) SL1344/pBR322_TIMER colors 1 or 4 days after i.v. infection (top panel), in various organs after appearance of disease signs in orally infected mice (middle panel), and in mesenteric lymph nodes (mLN) and spleen of streptomycin-pretreated mice at day 3 p.i. (bottom panel). The data represent pooled data for 3 to 4 mice from 2 to 3 experiments. The proportion of orange Salmonella with log(green/orange) < −0.055 (dashed lines) in individual mice is shown in (B). Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S3 Confocal Microscopy of TIMERbac-Salmonella-Infected Spleen, Related to Figure 4 (A) Analysis of Salmonella colors in single infected host phagocytes. Green/orange ratios of individual Salmonella are compared with averages colors of all Salmonella in the same cell. Most Salmonella had individual colors close to the average for the respective phagocyte. (B) Salmonella green/orange ratios in infected cells containing 1-2, or more Salmonella. (C) Salmonella green/orange ratios in infected F4/80hi resident macrophages and CD11bhi infiltrating host cells. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S4 Colors and Division Rates of Salmonella Expressing Low Levels of TIMERbac, Related to Figure 5 (A) Color distributions for SL1344/pSC101_TIMER in spleen of 129/Sv and BALB/c mice 4 days after i.v. infection with 700 CFU. For comparison, data for SL1344/pBR322_TIMER in BALB/c mice at the same time point are also shown (shaded area). Data were pooled for four mice from one experiment in each group. Median log(green/orange) ratios for SL1344/pSC101_TIMER in individual mice are shown in the inset. (B) Calculated division rate distributions based on data in (A) (averages ± SEMs for four mice from one experiment). The inset shows Salmonella loads in spleen of individual mice. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S5 Metabolism of Salmonella with Differential Growth Rates, Related to Figure 6 (A) Abundance ratios for metabolic enzymes based on proteome comparisons of green and orange TIMERbac-Salmonella subsets (see Table S1 for the full data set). Symbols represent metabolites (squares, carbohydrates; triangles, amino acids; circles, other metabolites; filled symbols, phosphorylated metabolites) and proteins (diamonds). The connecting lines present enzymes catalyzing the corresponding conversions. The brown lines represent the inner and outer membranes. An interactive version of this map with detailed descriptions for all reactions is available at http://www.biozentrum.unibas.ch/personal/bumann/Claudi_et_al/TIMER.html. (B) Analysis of in vivo division numbers of SL1344 and its purine-auxotrophic derivate purH using the Fluorescence Dilution technique. The gray areas show GFP fluorescence of arabinose-induced S1344/pFcGi and S1344 purH/pFcGi cultures used as inocula. The black lines show Salmonella GFP levels in spleen at day 2 postinfection. In both strains, most Salmonella diluted their GFP contents more than 20-fold (equivalent to > 4 divisions) during this infection time interval. However, a distinct subset in purH (but not SL1344) retained high GFP levels indicative of slow growth. The proportion of this subset is small compared to nongrowing subsets observed for purH using the TIMERbac approach (Figure 6B). This could reflect that in the Fluorescence Dilution approach only Salmonella with minimal growth throughout the entire infection period retain distinct high GFP levels. Such arrested Salmonella would be rapidly overgrown by normally proliferating Salmonella. Salmonella that initially divided a few times and diluted their GFP, but then entered microenvironments with poor purine supply and slowed down, might be indistinguishable from the majority of continuously growing Salmonella based on Fluorescence Dilution, whereas they would acquire detectably distinct orange color ratios in the TIMERbac approach. Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S6 Impact of Enrofloxacin Treatment Salmonella Survival and TIMERbac Color, Related to Figure 7 (A) Colony-forming units and fluorescent particle counts in chemostat cultures growing at 0.13 h-1, 1 hr after addition of 5 mg l-1 (final concentration) enrofloxacin or vehicle. Most treated Salmonella are killed but remain detectable as fluorescent particles with unaltered green/orange ratios (inset) at least for 1 hr after treatment. Each symbol represents one reactor from one experiment. (B) Salmonella CFU and fluorescent particle counts 1 hr after an i.p. dose of 0.2 mg enrofloxacin in spleen of i.v. infected mice (open red circles), an i.p. dose of 0.1 mg enrofloxacin in spleen of orally infected mice with clear disease signs during multi-dose therapy (filled red circles), or vehicle in i.v. infected mice (open black circles). Similar to in vitro conditions, Salmonella loose colony-forming capability but retain fluorescence. The inset shows median green/orange ratios in orally infected mice without treatment (black) or 1 hr after the first 0.1 mg enrofloxacin dose. Each symbol represents an individual mouse. (C) Proportions of Salmonella subsets with different division rate at day 4 after i.v. infection (averages ± SD of three mice from three independent experiments). (D) Salmonella survival (CFU per sorted TIMERbac-Salmonella cell) in subsets with different division rates in spleen of i.v. infected mice at day 4, 1 hr after an i.p. dose of 0.2 mg enrofloxacin. Each line represents data for one mouse from three independent experiments. Salmonella survival correlated with division rate (one-way repeated-measures ANOVA with posttest for linear trend). (E) Proportions of Salmonella subsets among survivors 1 hr after an i.p. dose of 0.2 mg enrofloxacin (averages ± SD of three mice from three independent experiments). Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions

Figure S7 Colonization of Various Mouse Organs at Early Time Points after Oral Infection, Related to Figure 7 Each symbol represents one mouse. Data were derived from unpublished data of a previous study (Bumann, 2002). Cell 2014 158, 722-733DOI: (10.1016/j.cell.2014.06.045) Copyright © 2014 Elsevier Inc. Terms and Conditions