Supplementary Figure 1. Analysis of LN S1P gradients.

Transcription

1 Supplementary Figure 1 Analysis of LN S1P gradients. (a) Diagram showing the spatial distribution of NK cells in WT LN. Inset: Upon infection, sinus-lining macrophages secrete IL-18 and other cytokines, which stimulate IFN- production by NK cells. IFN- in turn activates microbicidal activity in the macrophages. (b) Diagram showing the response to S1P by cells expressing the reporter construct. (c) FTY720 treatment demonstrates that reporting macrophages within the medulla are not saturated with S1P. S1P reporter mice (Mx1-Cre) were treated i.p. with 1 mg/kg of the S1PR1 agonist FTY720 or

2 vehicle 5h before LN harvest. LN sections were stained with antibodies against GFP (green), RFP (red), CD11b (white), Lyve1 (not shown) and B220 (not shown). Left: Representative cells from the medullary cords and medullary sinuses of FTY720-treated and vehicle-treated S1P reporter mice. Scale bar, 5 m. Representative of 2 pairs of mice analyzed in 2 experiments (image file: Right: Quantification of the GFP:RFP ratio on the surface of reporting cells as in Fig. 1c. Each point on the graph represents the average ratio over an area of at least 7x10 2 square microns. Horizontal lines indicate the mean and SEM. Ratios were not normalized. Graph shows data from 1 pair of mice, representative of 2 experiments. *, p<0.01; **, p< (Fisher s LSD test). (d) Reporter-transduced T cells sense more S1P in the medullary cords than the T zone. T cells isolated from WT CD or Cd69 / CD mice were retrovirally transduced with the S1P reporter and intravenously transferred to WT CD mice. LN were harvested for sectioning and confocal imaging 1d after transfer. We used Cd69 / T cells to avoid CD69-mediated S1PR1 internalization, although we did not see any differences between reporting by Cd69 / and WT T cells (not shown) 27. Left: LN sections were stained with antibodies against GFP (green), RFP (red), CD45.1 (to identify the surface of transferred cells, not shown), Lyve1 (not shown), and B220 (not shown). Representative cells are shown from the T zone and medullary cords. Scale bars, 5 m (image files: Right: Quantification of S1P reporting by transferred T cells in the T zone and medullary cords. The ratio of GFP to RFP for each CD RFP + pixel was calculated, and the ratio was averaged over individual cells. Graph compiles 19 cells in the medullary cords and 55 cells in the T zone from 16 sections (14 from 2 mice with Cd69 / donor T cells analyzed in 2 experiments, and 2 from 1 mouse with WT donor T cells in 1 experiment). Center line represents the median, box limits extend from the 25 th to 75 th percentile, Tukey whiskers extend to 1.5 IQR beyond the box, and symbols represent outliers. For each experiment all ratios were normalized such that the average ratio for the T zone was 1.0. *, p<0.01 (unpaired Student s 2- tailed t-test). (Supplementary Note 2)

3 Supplementary Figure 2 Identification of NK cells in tissue sections. Most NK1.1 + cells in LN sections are NKp46 + NK cells. Confocal immunofluorescence image of a LN section from a WT animal stained with anti-nk1.1 (green), anti-nkp46 (magenta), and anti-lyve1 (blue). Image is representative of 2 mice in 2 experiments. Scale bars, 20 m. More than 85% of NK1.1 + cells are also positive for NKp46. In most experiments, we identified NK cells by expression of NK1.1 rather than the more NK-specific NKp46 because the anti-nk1.1 antibody was more robust in our hands. Because the vast majority of cells that stain with NK1.1 in LN sections are NKp46 +, potential mislocalization of NKT cells or other NK1.1 + subsets would not substantially affect the results.

4

5 Supplementary Figure 3 NK cell defects are not accompanied by gross mislocalization of CD169 + macrophages. In WT mice, CD169 + macrophages line the SCS and medullary sinuses 32. Representative images of LN from the indicated animals showing localization of macrophages. B cells were stained with anti-b220, LEC and some sinus-lining macrophages with anti-lyve1, and a subset of macrophages with anti-cd169. Scale bars, 100 m. (a) LN sections from a Spns2 Lyve1 mouse (bottom) and littermate control (top). Representative of 2 pairs of mice in 2 experiments. (b) LN sections from an S1pr5 / mouse (bottom) and littermate control (top). Representative of 2 pairs of mice in 2 experiments. (c) LN sections from an AMD3100-treated mouse (bottom) and PBS-treated control (top). B cells (B220) and LEC (Lyve1) were stained with the same fluorophore but can be distinguished by morphology. Representative of 2 pairs of mice in 2 experiments. (d) LN sections from a Cxcr4 Mx1 mouse (bottom) and littermate control (top). Representative of 2 pairs of mice in 2 experiments.

6 Supplementary Figure 4 Cell-intrinsic requirement for S1PR5 in NK cell localization. (a) NK cell frequency in LN of Spns2 Lyve1 mice. NK cell frequency (left) and absolute number (right) in peripheral LN (2 inguinal, 2 axillary, 2 brachial) of Spns2 Lyve1 mice and littermate controls. As previously reported 26, 44, SPNS2-deficient mice have small LN compared to controls. Lines indicate the mean. Graphs compile 6 pairs of mice analyzed in 3 experiments. *, p<0.05 (unpaired Student s 2-tailed t-test). (b) NK cell frequency in LN of S1pr5 / mice. NK cell frequency (left) and absolute number (right) in peripheral LN (2 inguinal, 2 axillary, 2 brachial) of S1pr5 / mice and littermate controls. Lines indicate the mean. Graphs compile 6 (left) or 5 (right) pairs of mice analyzed in 4 (left) or 3 (right) experiments. *, p<0.05 (paired Student s 2-tailed t-test).

7 (c) Original images for Fig. 4f. LN sections were stained with antibodies against NK1.1 (magenta), Lyve1 (blue), and B220 (blue) (although they were both stained with BV421-conjugated antibodies, Lyve1 + LEC and macrophages and B220 + B cells can be readily distinguished by morphology). Note: the occasional cells expressing very bright GFP did not co-stain with NK1.1. Scale bars, 200 m. Insets: Green arrows indicate GFP + NK cells. White arrows indicate GFP - NK cells. Single channel images of GFP are shown at the bottom. Scale bars, 20 m.

8 Supplementary Figure 5 NK cell localization during infection. (a) NK cell frequency in LN of AMD3100-treated mice. NK cell frequency (left) and absolute number (right) in peripheral LN (2 inguinal, 2 axillary, 2 brachial) of AMD3100- or PBS-treatedmice. LN were harvested 2h after treatment. Lines indicate the mean. Graphs compile 6 pairs of mice analyzed in 2 experiments. (b) A subset of LN NK cells expresses CXCR3 during homeostasis. Surface CXCR3 staining on LN NK cells (NK1.1 + CD3 ) and CD4 + T cells in a representative Spns2 Lyve1 mouse and littermate control. Representative of 4 pairs of mice in 2 experiments.

9 macro "LN Reporter Regions" { //Draw 4 sets of ROI's and place them in the same folder as the image. //set GFP and RFP thresholds by the negative image var thresholdgfp = 16; var thresholdrfp = 23; var thresholdpe = 4; var thresholdf480 = 32; //get image attributes name = getinfo("image.filename"); original = getimageid(); originaltitle = gettitle(); from"); //Set the directory where files will be pulled from and saved to OpenDir=getDirectory("Choose the folder where data will be pulled DataDir=getDirectory("Choose the folder where data will be saved"); //convert to 32 bit run("32-bit"); // split channels and threshold the GFP and RFP images. //If GFP=0, the ratio image will be 0, and will become NaN during background/nan step //If RFP=0, the ratio will be x/0, which should return NaN, due to Misc options run("duplicate...", "title=gfpimage duplicate channels=1"); GFPImage = getimageid(); GFPImageTitle = gettitle(); changevalues(0, thresholdgfp, 0) selectimage(original); run("duplicate...", "title=rfpimage duplicate channels=2"); RFPImage = getimageid(); RFPImageTitle = gettitle(); changevalues(0, thresholdrfp, 0) // divide GFPImage by RFPImage run("misc...", "divide=nan"); imagecalculator("divide create 32-bit", GFPImage, RFPImage); rename("ratioimage"); RatioImage = getimageid(); RatioImageTitle = gettitle(); // generation of membrane mask selectimage(original); run("duplicate...", "title=pe duplicate channels=3"); //run("median...", "radius=1.0"); setthreshold(thresholdpe, 255); run("convert to Mask", " black"); PEImage = getimageid(); rename("memmaskimage") //to make the mask values either 1 or 0. run("divide...", "value= "); MemMaskImage = getimageid();

10 MemMaskTitle = gettitle(); // calculate RFP to GFP ratio on membrane. imagecalculator("multiply create 32-bit", RatioImage, MemMaskImage); setthreshold(0.0001, ); run("nan Background"); rename(originaltitle + "MemRatio"); MemRatioImage = getimageid(); MemRatioTitle = gettitle(); run("fire"); saveas("tiff", DataDir+MemRatioTitle+".tif"); setminandmax(0.0000, 4.000); // cleanup and select results table //selectimage(original); close(); selectimage(rfpimage); close(); selectimage(gfpimage); close(); //selectimage(ratioimage); close(); //Ratio measurements run("set Measurements...", "area mean display redirect=none decimal=3"); roimanager("reset"); open(opendir + "RoiSetMed.zip"); ROICountMed=roiManager("Count");//=3 for (i=0; i<roicountmed; i++) { roimanager("select",i); roimanager("rename", "Med"); //roimanager("reset"); open(opendir + "RoiSetMedL.zip"); ROICountMedL=roiManager("Count"); for (i=roicountmed; i<roicountmedl; i++) { roimanager("select",i); roimanager("rename", "MedL"); //roimanager("measure"); //roimanager("reset"); open(opendir + "RoiSetT.zip"); ROICountT=roiManager("Count") for (i=roicountmedl; i<roicountt; i++) { roimanager("select",i); roimanager("rename", "T"); open(opendir + "RoiSetSCS.zip"); ROICountSCS=roiManager("Count") for (i=roicountt; i<roicountscs; i++) { roimanager("select",i); roimanager("rename", "SCS"); roimanager("deselect"); roimanager("measure"); // save the results table as an excel file //Adding the settings to the results table: i = nresults; // variable for counting, initialising with 0 setresult("image name", i, originaltitle);

11 setresult("thresholdgfp", i, thresholdgfp); // add a "Label" column to the results table and name the entry "point1" setresult("thresholdrfp", i, thresholdrfp); setresult("thresholdpe", i, thresholdpe); setresult("thresholdf480", i, thresholdf480); selectwindow("results"); saveas("results", DataDir + originaltitle + ".xls");!

12 macro "LN Reporter transduced primary T cells" { var thresholdgfp = 5; var thresholdrfp = 30; var thresholdpe = 12; from"); //get image attributes name = getinfo("image.filename"); original = getimageid(); originaltitle = gettitle(); shorttitle= substring(originaltitle, 0,11) CellROI="Cells " + substring(originaltitle, 0,11) + ".zip"; MedROI="MedROI.zip"; TROIs="TROI.zip"; TROI="TROI.zip"; //Set the directory where files will be pulled from and saved to OpenDir=getDirectory("Choose the folder where data will be pulled DataDir=getDirectory("Choose the folder where data will be saved"); //convert to 32 bit run("32-bit"); //split channels and threshold the GFP and RFP images. //If GFP=0, the ratio image will be 0, and will become NaN during background-->nan //If RFP=0, the ratio will be x/0, which should return NaN, as set in Misc options run("duplicate...", "title=gfpimage duplicate channels=1"); GFPImage = getimageid(); GFPImageTitle = gettitle(); changevalues(0, thresholdgfp, 0) selectimage(original); run("duplicate...", "title=rfpimage duplicate channels=2"); RFPImage = getimageid(); RFPImageTitle = gettitle(); changevalues(0, thresholdrfp, 0) // divide image1 by image2 run("misc...", "divide=nan"); imagecalculator("divide create 32-bit", GFPImage, RFPImage); rename("ratioimage"); RatioImage = getimageid(); RatioImageTitle = gettitle(); // generation of membrane mask //Mask the PE selectimage(original); run("duplicate...", "title=pe duplicate channels=3"); run("median...", "radius=.5"); setthreshold(thresholdpe, 255); run("convert to Mask", " black"); PEImage = getimageid(); rename("memmaskimage") //to make the mask values either 1 or 0.

13 run("divide...", "value= "); MemMaskImage = getimageid(); MemMaskTitle = gettitle(); // calculate RFP to GFP ratio on membrane. imagecalculator("multiply create 32-bit", RatioImage, MemMaskImage); //run("threshold..."); setthreshold(0.0001, ); run("nan Background"); rename(originaltitle + "MemRatio"); MemRatioImage = getimageid(); MemRatioTitle = gettitle(); run("fire"); setminandmax(0.0000, 1.000); saveas("tiff", DataDir+MemRatioTitle+".tif"); roimanager("reset"); if (File.exists(OpenDir + CellROI) == 1) x=x; else { settool("rectangle"); waitforuser("draw the T cell ROIs as small as possible"); roimanager("save", DataDir + CellROI); if (File.exists(OpenDir + MedROI) == 1) x = x; else { roimanager("reset"); settool("polygon"); waitforuser("draw the Medulla ROIs and add to manager"); ROICount=roiManager("Count");//=3 for (i=0; i<roicount; i++) { roimanager("select",i); roimanager("rename", "Med"); roimanager("save", DataDir + MedROI); //here, open all the ROIs, which are named properly now. Then open the T zone ROI. if (File.exists(OpenDir + TROIs) == 1 File.exists(OpenDir + TROI)) x = x; else { roimanager("reset"); settool("polygon"); waitforuser("draw the T zone ROIs and add to manager"); ROICount=roiManager("Count");//=3 for (i=0; i<roicount; i++) { roimanager("select",i); roimanager("rename", "T"); roimanager("save", DataDir + TROI); //generate a single ROI from all the TROIs roimanager("reset"); open(datadir + TROIs); ROICount=roiManager("Count"); roimanager("deselect"); roimanager("combine"); roimanager("add");

14 roimanager("select",roicount); roimanager("rename", "Whole T"); a1 = newarray(roicount); for (i=0; i<a1.length; i++) a1[i] = i; roimanager("select", a1); roimanager("delete"); //rename the cells that have a centroid inside the T zone open(datadir + CellROI); ROICount=roiManager("Count"); CellStart=1; CellEnd=ROICount-1; for (i=1; i<roicount; i++) { roimanager("select", i); Roi.getBounds(x, y, width, height); a= x+width/2; b= y+height/2; roimanager("select", 0); if (Roi.contains(a,b)==1) { roimanager("select", i); roimanager("rename", "T zone cell"); //open the medulla ROIs and generate a single ROI. ROICount1=roiManager("Count"); //9 open(datadir + MedROI); ROICount2=roiManager("Count"); //11 array = Array.getSequence(ROICount2); Selection=Array.slice(array,ROICount1,ROICount2); roimanager("select", Selection); roimanager("combine"); roimanager("add"); roimanager("select",roicount2); roimanager("rename", "Whole Med"); //rename the cells that have a centroid inside the medulla for (i=1; i<roicount; i++) { roimanager("select", i); Roi.getBounds(x, y, width, height); a= x+width/2; b= y+height/2; roimanager("select", ROICount2); if (Roi.contains(a,b)==1) { roimanager("select", i); roimanager("rename", "Med cell"); ROICountTotal=roiManager("Count"); //11 arraytotal = Array.getSequence(ROICountTotal); SelectionFinal=Array.slice(arrayTotal,ROICount,ROICountTotal); selectwindow(memratiotitle+".tif"); roimanager("select", SelectionFinal); roimanager("delete"); roimanager("select", 0); //deleting the T zone ROI roimanager("delete"); //Measure the cell ratios

15 run("set Measurements...", "area mean display redirect=none decimal=3"); roimanager("deselect"); roimanager("measure"); i = nresults; // variable for counting, initialising with 0 setresult("image name", i, originaltitle); setresult("thresholdgfp", i, thresholdgfp); // add a "Label" column to the results table and name the entry "point1" setresult("thresholdrfp", i, thresholdrfp); setresult("thresholdpe", i, thresholdpe); selectwindow("results"); saveas("results", DataDir + shorttitle + ".xls");