Droplet Tn-Seq combines microfluidics with Tn-Seq for identifying complex single-cell phenotypes

While Tn-Seq is a powerful tool to determine genome-wide bacterial fitness in high-throughput, culturing transposon-mutant libraries in pools can mask community or other complex single-cell phenotypes. Droplet Tn-Seq (dTn-Seq) solves this problem by microfluidics facilitated encapsulation of individual transposon mutants into growth medium-in-oil droplets, thereby enabling isolated growth, free from the influence of the population. Here we describe and validate microfluidic chip design, production, encapsulation, and dTn-Seq sample preparation. We determine that 1–3% of mutants in Streptococcus pneumoniae have a different fitness when grown in isolation and show how dTn-Seq can help identify leads for gene function, including those involved in hyper-competence, processing of alpha-1-acid glycoprotein, sensitivity against the human leukocyte elastase and microcolony formation. Additionally, we show dTn-Seq compatibility with microscopy, FACS and investigations of bacterial cell-to-cell and bacteria-host cell interactions. dTn-Seq reduces costs and retains the advantages of Tn-Seq, while expanding the method’s original applicability.

. Moreover, we show that besides the ability to resolve complex single-cell 55 behavior, droplets have many more advantages and applications including a drastic reduction in culture 56 media volume (and possible expensive compounds), and analyses of interactions between bacteria 57 and/or host-cells.  To determine the functionality of dTn-Seq, transposon insertion libraries of Streptococcus pneumoniae 78 were grown in batch-culture as a pooled population ('standard' Tn-Seq) and encapsulated as single cells 79 (dTn-Seq) in growth media with either glucose or the complex glycan alpha-1-acid glycoprotein (AGP) 80 as the major carbon source. Moreover, by adding low-melting temperature agarose to growth media with 81 glucose, droplets with a 1% agarose density were generated to assess how a solid environment that 82 provides structural support affects single cell growth. For each gene in each condition, fitness (i.e. the 83 growth rate) was calculated and compared between pooled-batch and droplet conditions. Overall, 2-5% 84 of genes from a variety of categories, including metabolism, transport, regulation and cell wall integrity  To validate dTn-Seq a total of 7 genes were chosen from the three environments (Supplementary Fig. 1; 89 Supplementary Tables 1-5). In the simplest environment with glucose as the carbon source ΔlytB 90 (SPT_1238) has no effect on fitness, however when grown in droplets the mutant has a severe growth  Table 1). This defect seems to be due to the 92 small droplet environment, since poor growth is masked when the mutant is grown by itself in batch 93 culture (5ml; Fig. 2b). LytB is part of the lytic cycle of S. pneumoniae and is involved in cell-chain 94 shortening (García et al. 1999). Indeed, chain-length of ΔlytB is significantly longer then the wt when 95 grown in batch, however when the mutant is grown in droplets, chain-lengths are shortened and 96 5 indistinguishable from wt (Fig. 2a). Recently longer cell chains were associated with rapid local 97 induction of competence (Domenech et al. 2018), which we hypothesized could be further enhanced in 98 the micro-droplet environment. Gene-expression of a set of competence genes was compared between 99 wt and ΔlytB grown in batch and in droplets. As posited, competence genes of ΔlytB cultured in droplets 100 are highly upregulated (Fig. 2c). Importantly, competence also induces the autolysin cbpD as well as the 101 immunity gene comM. In ΔlytB-droplets comM is upregulated ~8-fold, while cbpD is upregulated ~28-102 fold. This means that, especially in a confined space, LytB is extremely important in limiting a local 103 hypercompetent phenotype, which when deleted triggers autolysis and fratricide and reduces fitness. We next compared transposon libraries grown in culture medium with AGP as the major carbon source.   To conclude, dTn-Seq is a valuable extension of Tn-Seq that is able to uncover novel single-cell 140 phenotypes associated with microenvironments, community factors, and solid environments that are 141 masked by Tn-Seq. dTn-Seq is applicable to practically any bacterium and any variation of Tn-Seq (e.g.    (MicroChem) using a spin coater (Laurell) and set by baking at 95°C. The photomask was aligned with 183 the silicon wafer and UV exposed followed by a post-exposure bake ramping from 65°C to 95°C over 4 184 minutes. The mold was developed using SU-8 developer (MicroChem) per the manufacturers guidelines 185 and rinsed with isopropanol and dH2O followed by a hardening bake from 100C to 200C over 5 minutes.

186
The PDMS chip was generated by mixing PDMS and curing agent (Dow Corning, Sylgard 184) in a 187 10:1 ratio and added to the mold, degassed with a vacuum, and polymerized at 65°C overnight.

188
Polymerized PDMS was cut from the mold and a biopsy punch (0.75mm -Shoney Scientific) was used 189 to create ports for tubing (PE-2 tubing -Intramedic). PDMS slabs were bonded to glass (Corning -2947, 190 75x50mm) at the clean room by washing the glass with acetone and isopropanol in a sonicator bath 191 while the PDMS was washed with isopropanol, followed by thorough drying with filtered nitrogen gas.

192
The channel side of the PDMS slab and the glass slide were treated with plasma (400sccm flow; 400 193 watts; 45 sec) using a faraday barrel screen. Plasma treated surfaces were quickly brought into contact 194 and pressed together and then placed at 65°C for 10 minutes to complete bonding. PicoSurf-1 oil to the 'oil inlet', cell culture to the 'aqueous inlet', and collect droplets from the 'droplet 204 outlet' (Fig. 1b). Cells were diluted based on droplet size and according to a Poisson distribution with 205 the goal of generating droplets that contained a single bacterial cell (Shapiro 2003). With our device a 206 concentration of 2.1 × 10 6 cells/ml encapsulated into ~157pL sized droplets will yield approximately 74% 207 empty droplets, 22% with single cells, and 3% with two or more cells. The syringe pump rate for cell 208 encapsulation was 400 µl hr -1 yielding ~1.4x10 6 total droplets in 30 minutes. To generate agarose 209 droplets the entire droplet production system was placed in a 37°C warm room. 1% Seaplaque agarose 210 was added to growth media and then heated until dissolved. The agarose was then filtered (0.22µm) after 211 which the cells were added to the agarose solution. After production, agarose droplets were gelled at 4°C            hardly grow when glucose is replaced by AGP (green bars). Importantly, this growth defect in AGP can 354 14 be resolved by adding wt to the culture (blue bars), indicating that wt is providing 'community support'.

355
(e) Agarose droplets can be generated by adding low melting agarose to growth media, which provides 356 structural support and results in bacteria (Gram-negative and positive alike) growing in microcolonies.