Massively Parallel Nanoblade Shows Francisella ... - Nature

4 downloads 0 Views 2MB Size Report
propidium iodide (red) was used to identify dead cells. Scale bar, 50 µm. (b) 200 nm magnetic beads (red) were delivered into HeLa cells and attracted by a ...
Supplementary Figure 1 Fabrication of the BLAST platform. The silicon-based device was fabricated on 2” double-sided polished silicon wafers of 300 μm thickness as follows: 1) A 1.5 μm thick thermal oxide layer was grown on both sides of a wafer. 2) After surface treatment with hexamethyldisilazane, photoresist AZ4620 (AZ Electronic Materials) was spin coated on one side of the wafer and exposed to UV light through a chrome mask with the design of fluid channels. The patterns were developed in dilute AZ400K (AZ Electronic Materials) solution for 5 min. 3) After 100°C baking for 10 min, the wafer was etched by advanced oxide etching (Surface Technology Systems) to remove the oxide layer and by deep reactive ion etching (Unaxis USA) to etch through the silicon substrate. 4) Another photolithography process was repeated on the opposite side of the wafer to pattern a 3 μm hole array and aligned to the channel structures. 5) Advanced oxide etching was used again to etch through holes on the oxide thin film. There was still a residue photoresist layer on the top surface. 6, 7) A 100 nm thick titanium thin film was deposited at a tilted angle by an e-beam evaporator (CHA Industries). 8) Finally, the device was immersed in acetone to detach the titanium thin film on the photoresist such that only the sidewalls of the small holes retained the metal coating.

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 2 Time-resolved images showing the dynamics of rapidly expanding bubbles triggered under different laser fluence and polarization (bubbles are the darker area of these images). Cavitation bubbles are initiated at the two tips of the tapered titanium film due to the lighting-rod enhancement effect near sharp metal

Nature Methods: doi:10.1038/nmeth.3357

tips. This is a non-resonant broadband effect and is also insensitive to light polarization on our platform. The bubble size is dependent on the laser fluence. Higher pulse energy generates larger sized cavitation bubbles with longer lifetimes. The lifetime of bubbles triggered by nanosecond laser pulses typically lasts for a few hundred nanoseconds. Scale bar, 3 μm.

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 3 40-kDa FITC-dextran delivered into HeLa and three types of primary human cells (3 d after delivery). (a-d) Bright-field microscopy images of (a) HeLa cells, (b) normal human dermal fibroblasts (NHDFs), (c) peripheral blood monocyte derived macrophages (PB-MDMs), and (d) renal proximal tubule epithelial cells (RPTECs). Scale bar, 50 μm. (e-h) 40kDa FITC-dextran (green) was delivered into (e) HeLa cells, (f) NHDFs, (g) PB-MDMs, and (h) RPTECs and incubated on chips for 3 d after delivery. Cells were stained by CellTracker™ Blue CMAC (blue, top row); propidium iodide (red) was used to identify the dead cells (bottom row). Scale bar, 100 μm.

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 4 Additional large cargo successfully delivered into cells with BLAST. (a) 2 μm green fluorescent polystyrene beads were delivered into HeLa cells. Cell membranes were stained with WGA 350 (blue), and propidium iodide (red) was used to identify dead cells. Scale bar, 50 μm. (b) 200 nm magnetic beads (red) were delivered into HeLa cells and attracted by a nearby micromagnet. Scale bars, 50 μm. (c) Mouse anti-α-tubulin monoclonal antibody with Alexa 488 conjugate was delivered into NHDFs, labeling the cytoskeleton (green). After 5 h incubation, cells were fixed and permeabilized to remove background fluorescence. Scale bar, 30 μm. (d) Delivery efficiency of 5 different sizes of polystyrene beads (20 nm, 200 nm, 500 nm, 1 μm, 2 μm) and cell viability. The data shows that the cell viability is similar and above 90% with delivery of all bead sizes. The delivery efficiency decreases with size from 93% for 20 nm beads to 62% for 2 µm beads. Error bars, s.d. (n = 4). (e) Co-delivery of 100 nm and 500 nm beads at a 6:1 ratio (180 beads/pl for 100 nm beads and 30 beads/pl for 500 nm beads). This ratio remains close to 6 after the beads were delivered into cells [92.4 ± 36.2 beads/cell for 100 nm beads (red) and 16.3 ± 7.4 beads/cell for 500 nm beads (green)], which proves that BLAST can deliver multiple payloads and the delivery process is not strongly size and diffusion speed dependent. Scale bar, 50 μm. (f) A control experiment shows that with fluid pumping but no laser pulsing, few 100 nm and 500 nm beads enter cells. Scale bar, 50 μm. (p < 0.0001 for beads/cell in e vs. f)

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 5 Evaluation of stress levels of HeLa cells after BLAST delivery. The HSPA6 gene expression level indicates the cellular response to heat shock. Gene expression levels were normalized to a GAPDH housekeeping gene for all cell samples including (A) laser pulsing with fluid pumping, (B) laser pulsing only, (C) fluid pumping only and (D) control without laser pulsing and fluid pumping. BLAST delivered cells (A-C) showed expression levels that were statistically similar to the cells without laser and flow applied (D). A positive control was provided by incubating HeLa cells in culture media at 42°C for 1 h. Error bars, s.d. (n = 3). Primers for gene expression were generated using Roche’s Universal Probe Library Assay Design Center online. HSPA6 Fwd: (TCATGAAGCCGAGCAGTACA) Rev: (GTTTTTGGCAGCCACTCTGT). GAPDH Fwd: (GCTCTCTGCTCCTCCTGTTC) Rev: (ACGACCAAATCCGTTGACTC).

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 6 Low-magnification data showing the uniformity of enzyme delivery. β-lactamase at 50 units/ml (a,c) and 1 unit/ml (b,d) in PBS was delivered into NHDFs using BLAST under the condition of fluid pumping with (a,b) or without (c,d) laser pulsing. The β-lactamase FRET substrate, CCF4 is added to the cells and under UV excitation fluoresces green in the absence of cytosolic β-lactamase (c,d). β-lactamase delivered to the cells by laser pulsing cleaves the CCF4, causing a loss of FRET and a shift in fluorescence from green to blue (a,b). Scale bar, 100 μm.

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 7 GFP-expressing parental L. monocytogenes, but not the hly plcB mutant strain, recruits actin and forms actin tails after phagocytic uptake by THP-1 cells. THP-1 cells were incubated with GFP-expressing parental or mutant strains L. monocytogenes at an MOI of 4:1 for 30 min in DMEM with 10% FBS, washed, and incubated in culture medium containing 5 g/ml gentamicin to kill extracellular bacteria. After 6 h, the monolayers were fixed with 4% paraformaldehyde in PBS and actin filaments were stained with phalloidin-AlexaFluor 549 and nuclei were stained with DAPI. GFP-fluorescent wild-type (a) and mutant (d) bacteria fluoresce green (shown in gray scale). Actin filaments stained by phalloidin-AlexaFluor 549 fluoresce red (shown in gray scale) are recruited to the parental L. monocytogenes (b) and frequently form comet tails (arrowheads), but no recruitment or comet tails are seen with the escape incompetent mutant strain (e). In marked contrast, as shown in Fig. 5a, this escape incompetent mutant strain recruits actin and forms abundant comet tails after direct cytosolic delivery by BLAST. Merged color images are shown in panels (c) and (f). Scale bars, 10 m.

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 8 Characterization of F. novicida mutants: growth in HeLa cells and macrophages and IglC expression. (a,b) Francisella Pathogenicity Island is required for cytosolic replication. Macrophages are the natural primary host cells for Francisella, and it is well established that Francisella is engulfed by macrophages into membrane-bound phagocytic vacuoles from which the bacterium must escape into the cytosol to replicate intracellularly. As a control, experiments in which THP-1 cells were infected with Francisella were carried out in parallel with BLAST experiments in which HeLa cells were infected with Francisella. F. novicida wild-type or F. novicida mutant with deletion of the entire Francisella Pathogenicity Island (FPI) were delivered into HeLa cells by BLAST (a) or ingested by phorbol ester-differentiated human macrophage-like THP-1 cells (b), as described. At 1, 6, and 24 h postinfection, the monolayers were lysed with 1% saponin in PBS and plated on chocolate agar for enumeration of bacterial colony forming units (CFU). While the wild-type F. novicida multiply in both HeLa and THP-1cells, the FPI strain is not able to multiply either in macrophages, in which it is unable to escape from the phagosome, or in HeLa cells, even when delivered directly into the cytosol. Error bars, s.d. (n = 3). (c) Complementation restores the ability of F. novicida iglC mutant to multiply intracellularly. Human monocytic THP1 cells (1×105/well) were differentiated with phorbol ester and infected with F. novicida expressing sfGFP (1×106/well) for 1.5, 6, or 24 h, as indicated. Three F. novicida strains were studied: F. novicida wild-type, F. novicida iglC, and F. novicida iglC complemented with iglC on a plasmid. The infected monolayers were then fixed with 4% formaldehyde, and the nuclei stained with DAPI. High resolution images of the infected monolayers were acquired by ImageXpress (Molecular Probes) and analyzed by MetaXpress High Content Image Analysis Software. From 1.5 to 24 h, the number of green fluorescent bacteria per nucleus increased about 10-fold for the F. novicida wild-type and complemented F. novicida iglC strains; in contrast, the number of F. novicida iglC per nucleus decreased by 3fold. Error bars, s.d. (n = 8, 3, and 8 wells for all strains at 1.5, 6, and 24 h, respectively). (d) Immunoblot analysis confirms the lack of 7 IglC expression in the F. novicida mutant strains. Bacterial lysates from 1.5×10 GFP expressing F. novicida wild type (lane 1), FPI (lane 2), iglC (lane 3), and iglC complemented with iglC on a plasmid (lane 4) strains were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membrane for probing with monoclonal antibody to IglB (1:1,000; BEI Resources) or polyclonal antibody to IglC (1:1,000), bacterioferritin (Bfr, 1:1,000), or GFP (1:2,000; Assay Designs), as indicated to the right of the figure. IglC expression is absent from F. novicida FPI and iglC strains (lanes 2 and 3) and restored in the iglC complemented strain (lane 4).

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Figure 9 Differential digitonin permeabilization assay for F. novicida localization in HeLa cells. GFP-expressing F. novicida wild-type strain (a-e), ΔiglC-mutant strain (f-j), or iglC-complemented strain (k-o) were delivered by BLAST into HeLa cells and the cells were fixed and stained by the differential digitonin method at 30 min (a-e), or 1 d (f-o) after laser assisted delivery. Cell membranes were stained with WGA-AlexaFluor 594 (near infrared, shown in gray scale in b,g,l). HeLa cells were permeabilized with digitonin (which permeabilizes plasma membrane but not intracellular vesicles) and antibody accessible F. novicida (i.e. those not enclosed within vacuoles) were stained with a chicken anti-F. novicida antibody followed by a goat anti-chicken Texas Red-X conjugated secondary antibody (red, d,i,n). GFP-expressing F. novicida fluoresce green (c,h,m) and HeLa cell nuclei are stained blue with DAPI (a,f,k). Merged images are shown in e, j and o. At 30 min after BLAST (a-e), bacteria stain both green and red,

Nature Methods: doi:10.1038/nmeth.3357

indicating cytosolic localization. By 1 d after BLAST, most of the ΔiglC mutant strain have been repackaged into vacuoles (f-j) and fluoresce green but not red (arrowheads), with only occasional bacteria fluorescing both green and red (arrow). In contrast, by 1 d after BLAST delivery, the iglC-complemented strain (k-o) has multiplied extensively within the HeLa cell and is stained by the red fluorescent antibody, confirming cytosolic localization in the HeLa cells. Scale bars, 20 μm.

 

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Note BLAST mechanistic information Cell membrane disruption using metallic nanoparticles under the illumination of short laser pulses has been widely studied1-3. On these metallic structures, the kinetic energy of oscillating electrons driven by electromagnetic waves quickly convert into lattice heat in picoseconds and starts to propagate into the surrounding aqueous media through thermal conduction4,5. Such rapid heating allows a substantial temperature rise in the nanostructure and the surrounding thin liquid layer over the laser pulse duration. Upon surpassing a threshold energy that superheats the liquid medium, part of the absorbed optical energy is converted into mechanical work through induction of explosive cavitation bubbles that generate highly localized and high-speed fluid flows. Depending on the pulse energy, pulse duration, and the metallic nanostructures, micron to nanoscale cavitation bubbles can be generated and shaped to trigger highly localized cell membrane disruption1. The degree of damage to cells induced by such a photothermal effect varies, depending on the size and the density of cavitation bubbles induced on a cell. Instant cell death occurs when there are too many bubbles induced or when the bubble size is too large. If the cavitation size and density is properly controlled, transient cell membrane permeability can be induced without killing the cells2,3,6. It should be noted is that if the laser pulse energy does not reach the threshold for triggering cavitation bubbles, the brief (tens of nanoseconds) transient temperature rise on nanostructures does not induce a significant effect on membrane permeability. In BLAST, the crescent shaped titanium thin films coated on the sidewall of each SiO2 hole aim to generate a cat-door-like cut to open a micron-sized membrane pore for large cargo delivery. The function of laser triggered cavitation bubbles is mainly to disrupt the cell membrane to create a pore for cargo delivery. The small volume perturbation (~ femtoliters) and random fluid flows induced by these bubbles play little role in introducing cargo into cells, especially for large-sized cargo. The main challenge for cargo delivery using BLAST is to achieve uniform delivery across the entire cell population on the chip. Modern microfabrication technology can ensure that the dimensions of these micro and nanostructures on BLAST are highly uniform across a 1 cm2 area, the current chip size. Laser pulsing needs to be done using a high repetition rate laser that scan the entire chip sufficiently rapidly such that the pores on the earlier scanned cells do not have time to reseal before allowing cargo delivery. Uniform and active fluid delivery into cells across these transient pores is the most critical and challenging part in the design. There are hundreds of thousands of SiO2 holes on a chip and tens of thousands of cells are randomly distributed. Holes fully covered or partially covered by cells have different flow resistances from holes not covered. How to ensure uniform delivery of cargo-carrying fluid into different cells is a major engineering challenge. The 3D microfluidic structures used in BLAST best address this challenge In the silicon structure, there are 10,000 vertical and short silicon channels [50 μm in diameter and 300 μm long (wafer thickness)] connecting the bottom cargo-storage reservoir and the SiO2 membrane hole array. These wide, short, vertical channels are to ensure that once the bottom elastic cargo storage chamber is squeezed, the fluid pressure can quickly and even distribute across the entire chip such that all

Nature Methods: doi:10.1038/nmeth.3357

SiO2 holes are under the same pressure and cargo carrying fluid is pushed into the cell cytosol through the transient pores that are opened. Fluid pumping needs to occur immediately after laser scanning is completed; otherwise, the membrane may reseal before cargo can be delivered into cells. The volume of fluid delivered needs to be experimentally optimized and calibrated. Large fluid deliveries cause immediate cell death and cell detachment. The fluid delivered along with desired cargo into a cell causes a small cell volume expansion. The elastic membrane then shrinks the cell volume back to its original size and shape. It is believed that some of the intracellular materials can also leak out through the transient membrane pores. On average, we estimate that 0.5 pl of fluid is delivered into cells, based on the number of beads delivered into a cell using a fluid with a known bead concentration (Supplementary Fig. 4e). Another major advantage of BLAST is it is “clog-free” delivery for cargo with sizes smaller than the diameter of the SiO2 membrane holes. We have never observed any clogging on the BLAST chip after numerous deliveries of all cargo types shown in this manuscript. This clog-free feature most likely can be attributed to the thin SiO2 membrane, which is only 1.5 μm thick, shorter than the diameter of the hole. Hence there is no structure to cause clogging as the cargo passes through the holes.

References 1. Lapotko, D. O., Lukianova, E. & Oraevsky, A. A. Selective laser nano-thermolysis of human leukemia cells with microbubbles generated around clusters of gold nanoparticles. Lasers Surg. Med. 38, 631-642 (2006). 2. Wu, T.-H. et al. Photothermal nanoblade for large cargo delivery into live mammalian cells. Anal. Chem. 83, 1321–1327 (2011). 3. Wu, T.-H. et al. Photothermal nanoblade for patterned membrane cutting. Opt. Express 18, 23153-23160 (2010). 4. Vogel, A. et al. Mechanisms of femtosecond laser nanosurgery of cells and tissues. Appl. Phys. B: Lasers Opt. 81, 1015-1047 (2005). 5. Kotaidis, V. & Plech, A. Cavitation dynamics on the nanoscale. Appl. Phys. Lett. 87, 213102 (2005). 6. Yao, C. et al. Elevation of plasma membrane permeability by laser irradiation of selectively bound nanoparticles. J. Biomed. Opt. 10, 064012 (2005).

Nature Methods: doi:10.1038/nmeth.3357

Supplementary Table 1 | Summary of all cell and cargo types delivered Cell Type

Delivered Cargo

Efficiency

Viability

HeLa

NHDF

Calcein Dextran (40 kDa) Gold particles (100 nm) Magnetic beads (200 nm) Francisella novicida β-Lactamase (29 kDa)

94.7±3.1 % 92.2±4.6 % 91.2±2.1 % 87.3±3.6 % 57.9±1.7 % 96.1±2.5 %

95.9±2.8 % † 97.2±1.5 % ‡ 95.3±1.9 % † 96.5±1.5 % † 95.9±1.2 % ‡ 97.5±0.9 % †

96.6±2.5 % 56.1±6.7 % 56.0±7.3 % 92.1±4.1 %

97.7±2.2 % ‡ ---

PB-MDM RPTEC

Dextran (40 kDa) Listeria Dextran (40 kDa) Dextran (40 kDa)

†: Propidium iodide exclusion assay 6 h after BLAST delivery ‡: Propidium iodide exclusion assay 24 h after BLAST delivery

Nature Methods: doi:10.1038/nmeth.3357

94.6±1.5 % ‡ 91.7±3.3 % ‡

Supplementary Table 2 | Strains, plasmids, and primers used in this study Strains, Plasmids, & Primers

Description

Source or Reference

Francisella tularensis subspecies novicida Fn Fn FPI

F. novicida Utah 112 strain Fn with the entire Francisella Pathogenicity Island deleted from the chromosome

Fn iglC

Fn with unmarked in-frame deletion of iglC

This study

sfGFP-Fn

Fn carrying pMP633-sfGFP

This study

sfGFP-Fn FPI

Fn FPI carrying pMP633-sfGFP

This study

sfGFP-Fn iglC iglC complemented sfGFP-Fn iglC

Fn iglC carrying pMP633-sfGFP Fn iglC carrying pMP633BC-iglC-sfGFP

This study This study

Listeria monocytogenes DP-L2318

GFP-L. monocytogenes Plasmids pMP590

Listeria monocytogenes hly plcB; 10403S with in-frame deletion of genes encoding the listeriolysin O and the broad-range phospholipase C DP-L2318 strain carrying pNF8

Ref. 7 Ref. 8

Ref. 9

This study

Suicide vector for allelic replacement in Francisella; sacB; KanR

Ref. 10

pMP590-iglC-ExC

pMP590 carrying an iglC gene deletion cassette containing the first 60 and the last 48 nucleotides of iglC and the ~1 kb chromosomal flanking regions

This study

pMP633 pMP633-sfGFP

Francisella shuttle plasmid; HygR sfGFP driven by bfr (FTL_0617) promoter inserted in MluI site on pMP633

Ref. 10 This study

pMP633BC-iglC-sfGFP

A bicistronic expression cassette of iglC and sfGFP driven by bfr (FTL_0617) and FTN_1451 promoter, respectively, inserted in the MluI site on pMP633

This study

pNF8

Listeria shuttle plasmid containing gfp-mut1; EmR

Primers

Nucleotide Sequence (5’ – 3’)

Fn ∆iglC screening ∆iglC-ExC-F

AGTGAGATGATAACAAGACAACAG

∆iglC-ExC-R

TTACTATGCAGCTGCAATATATCCTA

pMP590-∆iglC-ExC construction ∆iglC-ExC-Up-F

TCGGGATCCATGGGTATGGTGGCAAAGAA

∆iglC-ExC-Up-R

TATCTGTGCTAGCAGTTCTCACATGAATGGTCT

Nature Methods: doi:10.1038/nmeth.3357

Ref. 11

∆iglC-ExC-Dn-F

GAGAACTGCTAGCACAGATAAAGGAGTTGCT

∆iglC-ExC-Dn-R

GATCGCGGCCGCAAGCCGTAAAAACCGCACTA

pMP633-sfGFP construction Pbfr-Mlu-F1 PbfrSD-SnaB-R1 sfGFP-SnaB-F sfGFP-Mlu-SR

ATACGCGTGGTACCTGGTTACTATTGCCATCATCACA CCTTTACTCATTACGTACCTCCTATTGTTACCTCCATTATTTA AAACTC AGGAGGTACGTAATGAGTAAAGGTGAAGAGCTATTTACTG ATACGCGTGGATCCTCATTATTTATATAACTCATCCATTCCAT GAGT

pMP633BC-iglC-sfGFP construction Pbfr-Mlu-F2

AGCTTACGCGTTGGTTACTATTGCCATCATCACAATAT

PbfrSD-SnaB-R2

CATTACGTACCTCCTATTGTTACCTCCATTATTTAAAACTC

iglC-SnaB-F

Pomp-SacR

AGGAGGTACGTAATGAGTGAGATGATAACAAGACA TACAACCGGTCTCGAGCAATTGCTATGCAGCTGCAATATAT CCTA TCGAGACCGGTTGTACATTAATTAATTTTGGGTTGTCACTC ATCGTAT ACTCATTTTGAGCTCTCCTTTTTTTGTTATAAATATTTTAT

sfGFP-SacF

GAGAGCTCAAAATGAGTAAAGGTGAAGAGCTATTTACTG

iglC-Age-SR Pomp-AgeF

Nucleotides that are underlined indicate restriction sites.

References 7. Karl Klose, University of Texas San Antonio, San Antonio, TX 8. Weiss, D. S. et al. In vivo negative selection screen identifies genes required for Francisella virulence. Proc. Natl. Acad. Sci. U. S. A. 104, 6037-6042 (2007). 9. Marquis, H., Doshi, V., & Portnoy, D. A. The broad-range phospholipase C and a metalloprotease mediate listeriolysin O-independent escape of Listeria monocytogenes from a primary vacuole in human epithelial cells. Infect. Immun. 63, 4531-4534 (1995). 10. LoVullo, E. D., Sherrill, L. A., Perez, L. L., & Pavelka, M. S. Genetic tools for highly pathogenic Francisella tularensis subsp. tularensis. Microbiol. 152, 3425-3435 (2006). 11. Fortineau, N., Trieu-Cuot, P., Gaillot, O., Pellegrini, E., Berche, P., & Gaillard. Optimization of green fluorescent protein expression vectors for in vitro and in vivo detection of Listeria monocytogenes. Res. Microbiol. 151, 353-360 (2000).

Nature Methods: doi:10.1038/nmeth.3357