Cargo encapsulation in bacterial microcompartments: Methods and analysis

Abstract

Metabolic engineers seek to produce high-value products from inexpensive starting materials in a sustainable and cost-effective manner by using microbes as cellular factories. However, pathway development and optimization can be arduous tasks, complicated by pathway bottlenecks and toxicity. Pathway organization has emerged as a potential solution to these issues, and the use of protein- or DNA-based scaffolds has successfully increased the production of several industrially relevant compounds. These efforts demonstrate the usefulness of pathway colocalization and spatial organization for metabolic engineering applications. In particular, scaffolding within an enclosed, subcellular compartment shows great promise for pathway optimization, offering benefits such as increased local enzyme and substrate concentrations, sequestration of toxic or volatile intermediates, and alleviation of cofactor and resource competition with the host. Here, we describe the 1,2-propanediol utilization (Pdu) bacterial microcompartment (MCP) as an enclosed scaffold for pathway sequestration and organization. We first describe methods for controlling Pdu MCP formation, expressing and encapsulating heterologous cargo, and tuning cargo loading levels. We further describe assays for analyzing Pdu MCPs and assessing encapsulation levels. These methods will enable the repurposing of MCPs as tunable nanobioreactors for heterologous pathway encapsulation.

1. Introduction

Metabolic engineers have successfully employed scaffolding approaches to improve heterologous pathway performance through colocalization and spatial organization (Conrado et al., 2012; Delebecque, Lindner, Silver, & Aldaye, 2011; Dueber et al., 2009; Lee et al., 2013; Lin, Zhu, & Wheeldon, 2017; Moon, Dueber, Shiue, & Prather, 2010; Sachdeva, Garg, Godding, Way, & Silver, 2014). Several computational approaches further predict that scaffolding by compartmentalization would offer unique benefits for pathway performance enhancement (Conrado, Mansell, Varner, & DeLisa, 2007; Hinzpeter, Gerland, & Tostevin, 2017; Jakobson, Tullman-Ercek, & Mangan, 2018). Subcellular organization into discrete compartments, or organelles, is a trait that is normally associated with the eukaryotic domain of life. However, up to 16% of all bacterial genera contain genes encoding protein-based organelles, termed bacterial microcompartments (MCPs) (Abdul-Rahman, Petit, & Blanchard, 2013; Axen, Erbilgin, & Kerfeld, 2014; Erbilgin, McDonald, & Kerfeld, 2014).

MCPs vary in size (50–200nm) and are composed entirely of small protein subunits that self-assemble into an exterior shell around an interior enzymatic core (Chowdhury, Sinha, Chun, Yeates, & Bobik, 2014). This enzymatic core determines specific compartment function and varies between species. For example, RuBisCO is found at the core of carboxysomes, the cyanobacterial MCPs used for carbon fixation (Shively, Ball, Brown, & Saunders, 1973). Other MCPs encapsulate the metabolic pathways required for the breakdown of unique carbon sources such as ethanolamine or 1,2-propanediol (1,2-PD) (Axen et al., 2014; Bobik, Havemann, Busch, Williams, & Aldrich, 1999; Brinsmade, Paldon, & Escalante-Semerena, 2005; Chen, Andersson, & Roth, 1994; Kofoid, Rappleye, Stojiljkovic, & Roth, 1999; Stojiljkovic, B€aumler, & Heffron, 1995). These metabolic MCPs function primarily to sequester toxic or volatile intermediates in these degradation pathways and enable efficient utilization of the corresponding substrates (Havemann, Sampson, & Bobik, 2002; Huseby & Roth, 2013; Penrod & Roth, 2006; Sampson & Bobik, 2008).

The 1,2-propanediol utilization microcompartment (Pdu MCP) natively encapsulates the enzymes, substrates, and cofactors required to degrade 1,2-PD (Fig. 1). The components of the Pdu MCP are encoded on the 21-gene pdu operon, which is induced by 1,2-PD and the upstream positive transcriptional regulator PocR (Bobik, Ailion, & Roth, 1992; Bobik et al., 1999; Chen et al., 1994; Rondon & Escalante-Semerena, 1992, 1996). A multistep enzymatic pathway within the Pdu MCP metabolizes 1,2-PD to propionate (Bobik, Xu, Jeter, Otto, & Roth, 1997; Cheng, Fan, Sinha, & Bobik, 2012; Horswill & Escalante-Semerena, 1999; Leal, Havemann, & Bobik, 2003; Liu et al., 2007; Palacios, Starai, & Escalante-Semerena, 2003). In this process, propionaldehyde, a toxic intermediate, is produced and consumed while sequestered within the MCP (Sampson & Bobik, 2008). Recent studies indicate that mutations affecting the permeability of Pdu MCP shells lead to toxic effects, supporting that the Pdu MCP shell acts as a selective diffusion barrier, allowing the influx of 1,2-PD while preventing the efflux of propionaldehyde (Chowdhury et al., 2015; Crowley et al., 2010). Computational studies also demonstrate that sequestration of propionaldehyde within the Pdu MCP enables increased local concentrations of this intermediate, enhancing flux through the pathway (Jakobson, Tullman-Ercek, Slininger, & Mangan, 2017). In addition to metabolic enzymes, the Pdu MCP also encapsulates enzymes that recycle the essential cofactors coenzyme B₁₂, NAD+, and coenzyme A, creating private cofactor pools within the MCPs (Bobik et al., 1997; Chen et al., 1994; Cheng & Bobik, 2010; Cheng et al., 2012; Johnson, Buszko, & Bobik, 2004; Johnson et al., 2001; Leal et al., 2003; Liu, Jorda, Yeates, & Bobik, 2015; Liu et al., 2007).

Fig. 1.

The 1,2-propanediol utilization microcompartment (Pdu MCP). (A) Diagram of the Pdu MCP (hexagon) and its encapsulated pathway. The substrate 1,2-PD is metabolized to propionate through a multistep enzymatic pathway with propionaldehyde as a toxic intermediate. The enzymes, substrates, cofactors, and intermediates shown inside the hexagon are encapsulated within the Pdu MCP. Abbreviations: Ado-B₁₂, adenosylcobalamin; B₁₂(I), cob(I)alamin; CoA, coenzyme A; HO-B₁₂, hydroxycobalamin; NAD, nicotinamide adenine dinucleotide. (B) Transmission electron microscopy (TEM) image of purified Pdu MCPs (scale bar = 0.2μm). (C) TEM image of thin cell sections showing Pdu MCPs within S. enterica LT2 (scale bar = 200nm). Courtesy of Edward Y. Kim.

The native properties of the Pdu MCP make it a promising enclosed scaffold for heterologous pathway sequestration and organization. Recent efforts to engineer Pdu MCPs for this purpose have focused on controlling the encapsulation levels of heterologous cargo (Jakobson et al., 2016; Jakobson, Kim, Slininger, Chien, & Tullman-Ercek, 2015), altering the diffusion properties of the native Pdu MCP shell (Chowdhury et al., 2015; Slininger Lee, Jakobson, & Tullman-Ercek, 2017), and expressing Pdu MCPs in nonnative hosts (Graf, Wu, & Wilson, 2017; Parsons et al., 2008). In one notable study, Warren and colleagues developed an ethanol nanobioreactor in Escherichia coli by encapsulating two enzymes required for the production of ethanol from pyruvate in empty Pdu MCP shells composed of shell proteins from Citrobacter freundii (Lawrence et al., 2014). In this chapter, we describe methods used in our laboratory to express and characterize Pdu MCPs, with an emphasis on techniques that enable and assess heterologous cargo encapsulation. These methods include detailed protocols for expressing Pdu MCPs in the native Salmonella host, tuning and quantifying the relative cargo encapsulation levels, visualizing MCPs via fluorescence microscopy, and purifying MCPs for analysis and assessment using in vitro assays.

2. Practical considerations for cargo encapsulation

Repurposing MCPs for metabolic engineering applications requires the ability to control MCP expression, encapsulate heterologous cargo, and control the level of cargo loading; here, we discuss methods for each of these critical steps. All methods detailed here are written for work with Pdu MCPs from the native host Salmonella enterica serovar Typhimurium LT2 (hereafter referred to as Salmonella enterica LT2).

2.1. Expression of Pdu MCPs

The following protocol describes the induction and expression of Pdu MCPs in the native Salmonella enterica LT2 host strain. In this protocol, cells are grown in No Carbon-E (NCE) minimal media supplemented with succinate as a carbon source. The substrate 1,2-propanediol (1,2-PD) is also added to the growth media to induce the formation of Pdu MCPs.

2.1.1. Equipment

• Shaking incubator

• Glass shake flasks

• Pipettes

• Autoclave or sterile filtration equipment

• NanoDrop or other spectrophotometer

2.1.2. Media, reagents, and consumables

• Lysogeny Broth, Miller (LB-Miller, 10g/L tryptone, 5g/L yeast extract, 10g/L NaCl)

• LB-Miller agar plates (15g agar dissolved in 1L LB-Miller, 100mm×15mm Petri dishes)

• NCE Media: 29mM KH₂PO₄, 34mM K₂HPO₄, 17mM Na(NH₄) HPO₄, 1mM MgSO₄, and 50μM ferric citrate

  • Supplement with 42mM succinate and 55mM 1,2-PD
  • KH₂PO₄, K₂HPO₄, and Na(NH₄)HPO₄ are prepared and stored as a 50 stock, which is autoclaved
  • MgSO₄ and ferric citrate are prepared and stored as 1000×stocks, succinate is prepared and stored as a 1.68M stock, and 1,2-PD is prepared and stored as a 50% stock (v/v); each of these stocks is sterile filtered

2.1.3. Procedure

1. From a glycerol stock, streak Salmonella enterica LT2 onto an agar plate and incubate at 37°C overnight (12–16h)

2. Pick a single colony using a sterile toothpick or pipette tip and inoculate in 5mL LB-Miller

3. Grow the 5mL starter culture at 30°C (225 RPM) for 24h

   1. At these conditions, the OD₆₀₀ of the starter culture should be~3.5–4 after 24h4.
4. During the growth of the 5mL starter culture, prepare the required volume of NCE mediai.

   1. For most applications, we find that a 200mL culture in a 1L flask works well, but the culture volume can be scaled up or down as needed for the application
   2. Autoclave ddH₂O in the growth flask and add sterile, concentrated stocks of the other components once the water has cooled to room temperature

5. After 24h of growth, subculture at a dilution of 1:1000 into the NCE media

6. Grow at 37°C (225 RPM) until the culture reaches a final OD₆₀₀ between 1.0 and 1.57.

7. Carry out appropriate assays (described in Section 3) to assess the formation of MCPs

2.1.4. Notes

• Overexpression of the transcriptional regulator PocR, even in the absence of 1,2-PD, is also capable of inducing Pdu MCP expression and formation (Kim, Jakobson, & Tullman-Ercek, 2014).

• We observe that growth to a final OD₆₀₀ between 1.0 and 1.5 typically takes 18–20h.

• Functional Pdu MCPs or empty Pdu MCP shells have also been heterologously expressed in Escherichia coli and other Gram-negative bacteria, and these systems could provide alternative host options for different applications (Graf et al., 2017; Parsons et al., 2008, 2010; Wagner, Capitain, Richter, Nessling, & Mampel, 2017). If using a host other than the native Salmonella enterica LT2 strain, growth and induction conditions might require additional optimization.

2.2. Coexpression and encapsulation of heterologous cargo

In the native Pdu MCP system, enzymes are targeted for encapsulation by short N-terminal peptides, termed signal sequences. Several native Pdu signal sequences are known: the first 18–20 amino acids of the diol dehydratase medium subunit PduD (ssPduD), the first 20 amino acids of the phosphotransacylase PduL (ssPduL), and the first 18 amino acids of the propionaldehyde dehydrogenase PduP (ssPduP) (Fan & Bobik, 2011; Fan et al., 2010; Liu et al., 2015). These signal sequences are necessary for the encapsulation of their corresponding enzymes within the Pdu MCP but are also sufficient for the encapsulation of heterologous cargo (Jakobson et al., 2016, 2015; Lawrence et al., 2014; Wagner et al., 2017). To mediate the encapsulation of a cargo of interest, we append one of these signal sequences to its N-terminus. This tagged cargo can then be expressed from a plasmid in parallel with Pdu MCP expression.

2.2.1. Equipment

• See Section 2.1.1

2.2.2. Media, reagents, and consumables

• LB-Miller

• NCE Media supplemented with 42mM succinate, 55mM 1,2-PD, and appropriate concentration of antibiotic(s), as necessary

  • Antibiotics are used at half the concentration used in standard growth media

2.2.3. Procedure

1. From a glycerol stock, streak the expression strain onto an agar plate with the appropriate antibiotic(s), as necessary, and incubate at 37°C overnight (12–16h)

2. Pick a single colony using a sterile toothpick or pipette tip and inoculate in 5mL LB-Miller with appropriate antibiotic(s), as necessary

3. Grow the 5mL starter culture at 30°C (225 RPM) for 24hi.

   1. At these conditions, the OD₆₀₀ of the starter culture should be~3.5–4 after 24h

4. During the growth of the 5mL starter culture, prepare the required volume of NCE media, as described in Section 2.1.3

5. After 24h of growth, subculture at a dilution of 1:1000 into the NCE media

6. Grow the NCE cultures at 37°C (225 RPM

7. After growing the NCE cultures for 14–15h, check the OD₆₀₀

8. When the culture reaches an OD₆₀₀ between 0.4 and 0.6, add the inducer for the cargo expression plasmid at the appropriate concentration

9. Grow for an additional 4–6h postinduction, targeting a final OD₆₀₀ between 1.0 and 1.5

10. Carry out appropriate assays (described in Section 3) to assess the formation of MCPs and the expression and encapsulation of cargo

2.2.4. Notes

• Selecting an appropriate cargo expression system depends on the project goal: origins of replication, antibiotic resistance cassettes, promoter strengths, and promoter compatibility should all be carefully considered.

• While the native Pdu signal sequences ssPduD and ssPduP are the most widely used, other native, nonnative, and de novo signal sequences can also be used for encapsulation. We have successfully repurposed signal sequences from other MCP systems, including enzymes from the ethanolamine utilization (Eut) MCP and glycyl radical-generating protein MCP, to encapsulate cargo within Pdu MCPs (Jakobson et al., 2015). We also generated and utilized de novo signal sequences, which conformed to an identified motif shared by native signal sequences across MCP systems (Jakobson et al., 2015; Jakobson, Lee, & Tullman-Ercek, 2017). These signal sequences can be screened in combination with heterologously expressed cargos to find optimal expression and encapsulation levels for the intended application.

• Growthconditionsandinductiontimesmayrequireoptimizationforanew cargo of interest. For example, some enzymes may require supplemented growth media or different temperatures to achieve optimal activity, which may in turn alter the delineated growth conditions and induction times.

2.3. Control of encapsulation levels via induction timing

Beyond encapsulation, control of cargo loading levels is essential for the use of Pdu MCPs with heterologous enzymes. Our lab recently established that the levels of cargo encapsulation within Pdu MCPs can be modulated by adjusting the relative timing of MCP formation and cargo expression (Jakobson et al., 2016). We expressed the transcriptional regulator PocR from an anhydrous tetracycline-inducible (pTet) plasmid and various signal sequence-GFP reporters from arabinose-inducible (pBAD) plasmids. The induction time for pdu operon expression and MCP formation (via the expression of PocR and addition of 1,2-PD) was varied relative to the induction time for reporter expression, resulting in different levels of encapsulation (Fig. 2). The specific method used for this study is described here, but it should be noted that the overall workflow would be generalizable to other expression vectors and cargos of interest.

Fig. 2.

Control of encapsulation via induction timing. (A) Diagram of the method for controlling cargo encapsulation levels by modulating induction times. MCP formation (top, hexagons) is induced at different times relative to cargo expression (bottom), resulting in different levels of encapsulation. (B) Heat maps of normalized GFP fluorescence resulting from the expression of the encapsulation reporters ssPduD-GFP-ssrA (left) or ssPduP-GFP-ssrA (right) for various induction time combinations. At the indicated induction times, MCP formation was induced with 1,2-PD and aTc and cargo expression was induced with arabinose. These induction times are relative to when cultures typically reach an OD₆₀₀ of 0.4. Each pair of relative induction times for MCP formation vs cargo expression was evaluated. In each square, the top value indicates the arithmetic mean and the bottom value indicates the 95% confidence interval for measurements from three or more replicates. This panel is reproduced from Jakobson, C. M., Chen, Y., Slininger, M. F., Valdivia, E., Kim, E. Y., & Tullman-Ercek, D. (2016). Tuning the catalytic activity of subcellular nanoreactors. Journal of Molecular Biology, 428(15), 2989–2996. https://doi.org/10.1016/j.jmb.2016.07.006.

2.3.1. Equipment

• See Section 2.1.1

• 24-well blocks

2.3.2. Media, reagents, and consumables

• LB-Miller

• NCE Media supplemented with 42mM succinate and appropriate concentrations of antibiotic(s), as necessary

• Inducers: 1,2-PD, arabinose, and anhydrous tetracycline (aTc)

  • Arabinose and aTc are prepared and stored as sterile-filtered 1000 stocks (20% (w/v) and 1μg/mL, respectively)

2.3.3. Procedure

1. Follow steps 1–6 from Section 2.2.3

   1. Here, NCE cultures can be grown in 24-well block instead of glass shake flasks, with the prepared media aliquoted at 5mL per well. This allows for efficient screening of a wider array of induction conditions per experiment, while still providing enough sample for certain downstream analyses. Once induction timing is optimized, we recommend repeating select experiments using a 200mL culture volume, as described in Section 2.1.
2. After growing the NCE cultures for 11–12h, begin induction series

   1. As a starting point, MCP formation and cargo expression should be separately induced at 1-h time intervals over 3h; each pair of induction times for MCP formation vs cargo expression will need to be evaluated (see Fig. 2B)
   2. For MCP formation, add 1,2-PD to a final concentration of 55mM and induce PocR expression with 1ng/mL aTc
   3. For cargo expression, induce with 0.02% arabinose

3. Grow for an additional 5.5h once induction is complete for all samples

4. Carry out appropriate assays (described in Section 3) to assess the formation of MCPs and the expression and encapsulation of cargo

5. Repeat steps 1 through 4 using additional induction time intervals (see step 2i) as necessary

2.3.4. Notes

• For other expression vectors and cargos of interest, the induction series can be altered to span different time frames or combinations of induction times to optimize for the intended application.

• The fluorescent reporter constructs used in our original study were also appended with a C-terminal SsrA degradation tag; this is further explained in Section 3 and would not necessarily be required for all cargos of interest.

2.4. Control of relative encapsulation levels of multiple cargo

Tuning the relative levels of multiple cargo can help to balance reaction stoichiometry and enhance flux through an encapsulated pathway. Our lab established that the relative encapsulation levels of various cargo can be modulated by tuning the expression levels of individual components or by using different combinations of signal sequences (Jakobson et al., 2015) (Fig. 3). We used a fluorescence-based assay to measure the relative encapsulation levels of two fluorescent proteins, GFP and mCherry, tagged with different signal sequences. Each construct was encoded on an arabinose- or anhydrous tetracycline-inducible plasmid, and expression levels were tuned using various concentrations of the corresponding inducer, resulting in varying levels of encapsulation. As in Section 2.3, the specific method used for this study is described here, but the overall workflow would be generalizable to other expression vectors and cargos of interest.

Fig. 3.

Diagram of the method for controlling relative encapsulation levels using different cargo expression levels and signal sequence combinations. Different signal sequences (blue vs orange) are appended to different cargo (solid green vs hatched red). The individual expression levels of these constructs are varied, resulting in different relative encapsulation levels.

2.4.1. Equipment

• See Section 2.3.1

• Plate reader with fluorescence capability (we use a BioTek Synergy HTX multimode plate reader)

2.4.2. Media, reagents, and consumables

• LB-Miller

• NCE Media supplemented with 42mM succinate, 55mM 1,2-PD, and appropriate concentration of antibiotic(s), as necessary

• 96-well plates

• Phosphate-buffered saline (PBS)

• Kanamycin

• Arabinose and anhydrous tetracycline (aTc), prepared and stored as 1000×stocks

2.4.3. Procedure

1. Follow steps 1–7 from Section 2.2.3

   1. Here, as in Section 2.3.3, NCE cultures can be grown in 24-well blocks
2. When the cultures reach an OD₆₀₀ of 0.4, add the inducers for the cargo expression plasmids at the appropriate concentrationsi.

   1. As a starting point, induce the different signal sequence-GFP reporters with 0.0025% or 0.005% arabinose (w/v), and induce the different signal sequence-mCherry reporters with 0.5, 1, 10, or 25ng/mL aTc; uninduced cultures can also be tested

3. Grow cultures for 5.5h postinduction

4. In a 96-well plate, dilute cells 1:4 in PBS with 2g/L kanamycin (PBSKan) to halt further translation

5. For each sample, measure OD₆₀₀ and fluorescence for GFP (excitation: 485/20, emission: 516/20) and mCherry (excitation: 560/40, emission: 620/15) using a plate reader

6. For each sample, normalize the fluorescence measurements by OD₆₀₀ to assess the relative encapsulation levels for the different fluorescent reporters

2.4.4. Notes

• A flow cytometer could be used instead of a plate reader for fluorescence measurements; see Section 3.2 for this method.

• The fluorescent reporter constructs used in our original study were also appended with a C-terminal SsrA degradation tag; this is further explained in Section 3 and would not necessarily be required for all cargos of interest.

• If the cargos of interest are not fluorescent proteins, induction can be carried out as described earlier, but other assays (see Section 3) would be required to assess the encapsulation levels.

3. Assays for microcompartment analysis

Here, we describe several in vivo and in vitro assays used to evaluate MCP formation, cargo encapsulation, and cargo activity; we also outline a method for purifying Pdu MCPs. It should be noted that several of these assays require fluorescent reporters that are tagged with a C-terminal SsrA degradation tag in addition to the N-terminal signal sequence. The SsrA tag functions natively to target proteins for degradation by the ClpXP pro-tease (Farrell, Grossman, & Sauer, 2005; Gottesman, Roche, Zhou, & Sauer, 1998). When using these tags, we find that unencapsulated cargo is degraded in the cytoplasm while encapsulated cargo is protected from degradation by the Pdu MCP shell.

3.1. Fluorescence microscopy

When using fluorescent reporters as cargo, fluorescence microscopy can be used to visualize Pdu MCPs within the cell and evaluate encapsulation. We typically use ssPduD-GFP or ssPduD-GFP-ssrA as our fluorescent encapsulation reporter, but this method has proven generalizable to other combinations of signal sequences and fluorescent proteins, as well as enzyme-fluorescent protein fusions. Fluorescent puncta are observed when the reporter is successfully encapsulated within the MCPs (Fig. 4A and C), while diffuse fluorescence is observed for unencapsulated reporter (Fig. 4B and E). Diffuse fluorescence can be reduced by utilizing a reporter appended to a degradation tag (Fig. 4D and F).

Fig. 4.

Phase contrast and fluorescence microscopy of S. enterica LT2 expressing different GFP reporters. Cells expressing the encapsulation reporters ssPduP-GFP or ssPduP-GFP-ssrA are shown in the presence (A and C, respectively) and absence (B and D, respectively) of 1,2-PD. In the presence of 1,2-PD, MCPs form and puncta are observed, indicating reporter encapsulation. In the absence of 1,2-PD, MCPs do not form, resulting in diffuse fluorescence when expressing ssPduP-GFP (B) and minimal fluorescence when expressing ssPduP-GFP-ssrA (D); the SsrA tag confers degradation of unencapsulated reporter in the cytoplasm. Diffuse fluorescence is also observed when expressing GFP without tags (E), while minimal fluorescence is observed when expressing GFP-ssrA (F). This figure is adapted from Kim, E. Y., & Tullman-Ercek, D. (2014). A rapid flow cytometry assay for the relative quantification of protein encapsulation into bacterial microcompartments. Biotechnology Journal, 9(3), 348–354. https://doi.org/10.1002/biot.201300391.

3.1.1. Equipment

• Nikon Ni-U upright fluorescence microscope with a 100×oil immersion objective

• Andor Clara-Lite digital camera

• Nikon NIS Elements software

• Appropriate filter for fluorescent reporter of interest

• ImageJ or Adobe Photoshop software

3.1.2. Media, reagents, and consumables

• 3×1 inch microscope slides and 22 22mm cover glass

• Electron Microscopy Sciences Immersion Oil Type LDF

• Optical Lens Cleaner

3.1.3. Procedure

1. Grow and induce cultures as described in Section 2.2.3 with the fluorescent encapsulation reporter as the coexpressed cargo

2. Add 2μL of culture to a microscope slide and carefully cover the sample with cover glass while avoiding bubbles; the sample should spread evenly and completely across the area of the cover glass

3. Add a small drop of immersion oil to the top of the cover glass at the center of the sample area

4. Secure the prepared slide on the microscope stage and raise the stage until the objective is just touching the drop of immersion oil

5. Use the fine adjustment for the microscope stage to bring the cells into focus

6. Using the attached camera and NIS Elements software, capture a phase contrast image of the cells

7. Switch to the proper filter and software setting for the fluorescent reporter

8. Capture a fluorescence image of the same cells captured in the phase contrast image

   1. Make sure not to move the microscope stage between the phase contrast and fluorescence images; this will allow for direct comparisons and overlays of the images during analysis

9. Take additional images, as necessary, by moving the stage to focus the objective on a new area of the sample and repeating steps 5–8

10. Repeat steps 2–9 for each culture sample, as necessary

11. Clean the immersion oil from the objective using lens cleaner and shut down the instruments

12. Use ImageJ or Adobe Photoshop for image correction and analysis

3.1.4. Notes

• We typically take 2–3 images in each quadrant of the sample to ensure coverage while also making sure to avoid areas where images have already been captured.

• Images within the same data set should use the same exposure time when taking fluorescence images. We find that 200–400ms is sufficient for GFP and mCherry, but this can be optimized for different fluorescent proteins and encapsulation levels.

• If image adjustments are made during analysis (e.g., brightness, contrast, cropping), all images within the same data set should be adjusted identically.

3.2. Flow cytometry

Relative encapsulation levels of fluorescent cargo can be quantified using a flow cytometry assay developed in our lab (Kim & Tullman-Ercek, 2014). This assay links the encapsulation level of fluorescent cargo to measurements of cellular fluorescence, enabling rapid screening of samples in vivo. This assay requires that a C-terminal SsrA degradation tag is appended to cargo intended for encapsulation. Without the degradation tag, unencapsulated cargo remains diffuse within the cell, and it is not possible to differentiate between encapsulated and unencapsulated cargo using measurements of cellular fluorescence (Fig. 5A). With the degradation tag, unencapsulated cargo is degraded, and measurements of cellular fluorescence thus report on only the levels of encapsulated cargo (Fig. 5B). This assay also uses the selective induction of Pdu MCP formation to correct for background fluorescence or inefficient degradation. As before, Pdu MCP formation can be induced by the addition of 1,2-PD. When Pdu MCP formation is not induced, no cargo can be encapsulated, so all cargo is degraded (Fig. 5B). Any remaining fluorescence in this negative control can then be used to correct for background fluorescence in the induced sample.

Fig. 5.

Diagram of the use of degradation tags for the flow cytometry assay. (A) For cells expressing a GFP encapsulation reporter without a degradation tag (ss-GFP), unencapsulated GFP is not degraded and measurements of cellular fluorescence cannot differentiate between encapsulated GFP and diffuse GFP. (B) For cells expressing a GFP encapsulation reporter with a degradation tag (ss-GFP-ssrA), unencapsulated GFP is degraded in the cytoplasm, while encapsulated GFP is protected. Measurements of cellular fluorescence are therefore indicative of the levels of encapsulated GFP only. A negative control in which Pdu MCP formation is not induced is also used to correct for background fluorescence.

3.2.1. Equipment

• Flow cytometer (we have used both the Millipore Guava easyCyte 5HT flow cytometer and the Attune NxT Acoustic Focusing flow cytometer)

• FlowJo or other flow cytometry analysis software

• 24-well blocks

• Multichannel pipette

3.2.2. Media, reagents, and consumables

• LB-Miller

• NCE Media supplemented with 42mM succinate and appropriate concentration of antibiotic(s), as necessary; prepare NCE media both with and without 55mM 1,2-PD

• PBS-Kan

• 96-well plates

• Appropriate buffers and solutions for the specific flow cytometer

3.2.3. Procedure

1. Follow growth and induction steps as outlined in Section 2.2

   1. Subculture each sample into NCE media both with and without 1,2-PD
   2. Here, as in Sections 2.3 and 2.4, NCE cultures can be grown in 24-well blocks or flasks
2. After growth, dilute cells in PBS-Kan to the density recommended for the flow cytometer. Use a 96-well plate and multichannel pipette to efficiently dilute multiple samples simultaneously

   1. We find dilutions between 1:1000 and 1:4000 in 200μL of PBS-Kan to be adequate depending on the flow cytometer being used; check manufacturer’s recommendations

3. For each sample, collect 5,000–10,000 events on the flow cytometer

4. Use FlowJo or other flow cytometry analysis software to gate samples and assess fluorescence

   1. We use the forward and side scatter channels to gate events and distinguish cells from debris
   2. We use the geometric mean to calculate the average fluorescence values for each sample

3.2.4. Notes

• This assay was originally developed to assess the encapsulation levels mediated by different N-terminal signal sequences, but these techniques can also be used to assess encapsulation resulting from different growth conditions, induction conditions, or other experimental design considerations.

• The number of events to be collected can be varied to ensure that there are a sufficient number of events to derive statistically significant data after gating.

• If there are issues with cargo expression or aggregation, this assay should not be used. Fluorescence microscopy (see Section 3.1) can be used in conjunction with this assay to confirm encapsulation of fluorescent cargo and lack of aggregation.

3.3. Microcompartment purification

Intact Pdu MCPs can be purified from cells using a multistep centrifugation process (Fig. 6). The protocol outlined here is adapted from the MCP purification method originally developed by Sinha, Cheng, Fan, and Bobik (2012). For this process, cultures grown under the desired conditions are harvested and lysed. The lysed cells are then centrifuged to remove cell debris, with the MCPs remaining in the supernatant. The supernatant is further centrifuged to obtain a pellet containing the MCPs. This pellet is then collected and the purified MCPs assessed.

Fig. 6.

Diagram of the workflow for Pdu MCP purification. Harvested cells are lysed and MCPs are separated from cell debris using a series of centrifugation steps, resulting in a pellet containing purified Pdu MCPs.

3.3.1. Equipment

• See Section 2.1.1

• Centrifuge (we use a Beckman Coulter Avanti JXN-30 High Performance Centrifuge)

• Fixed-angle and swinging-bucket rotors (we use a FIBERLite F10BCI-6×500y fixed-angle rotor and a Beckman Coulter JS-24.38 swinging-bucket rotor)

• Nalgene PPCO Centrifuge Bottles (capacity 500mL)

• Microcentrifuge

• Rocker

3.3.2. Media, reagents, and consumables

• LB-Miller

• NCE Media supplemented with 42mM succinate, 55mM 1,2-PD, and appropriate concentration of antibiotic(s), as necessary

• Buffer A: 50mM Tris–HCl, 500mM KCl, 12.5mM MgCl₂, 1.5% 1,2-PD; pH to 8.0 using KOH

• Buffer B: 50mM Tris–HCl, 50mM KCl, 5mM MgCl₂, 1% 1,2-PD; pH to 8.0 using KOH

• Tris-OTG Buffer: 20mM Tris, 1% octylthioglucoside (OTG); pH to 7.5 using HCl

• Lysis Buffer (per sample): 5mL Buffer A, 7.5mL Tris-OTG Buffer, 10mg lysozyme, 0.25μL DNase I, 5mM β-mercaptoethanol

• Serological pipettes

• 50mL conical tubes

• Nalgene High-Speed Polycarbonate Round-Bottom Centrifuge Tubes (capacity 38mL)

• 1.5mL microcentrifuge tubes

3.3.3. Procedure

1. Follow growth and induction steps as outlined in Section 2.2

   1. We find that a 200mL culture in a 1L flask provides sufficient MCP yield for downstream analyses; volumes used in subsequent steps for lysis, washing, and collection are based on this culture volume, but can be scaled as necessary

2. Prepare Lysis Buffer for the appropriate number of samples and store at 4°C until use

3. When the target OD₆₀₀ of 1.0–1.5 is reached, transfer the culture into a centrifuge bottle

4. Harvest the culture at 5000 RPM for 5min using the centrifuge with the fixed-angle rotor

5. Decant the supernatant and resuspend the cell pellet in 12.5mL Lysis Buffer

6. Once fully resuspended in Lysis Buffer, transfer the sample into a 50mL conical tube

7. Incubate the sample at room temperature for 30min on a rocker set to 60 RPM

8. Incubate the sample on ice for 5min

9. Using the fixed-angle rotor and an appropriate adapter for the conical tube, centrifuge the sample at 12,000×g for 5min at 4°C to pellet cell debris

   1. Note that the MCPs will remain in the supernatant

10. Transfer the supernatant into a clean 50mL conical tube, taking care not to disturb the pellet of cell debris, and repeat step 9

11. Using a serological pipette, carefully transfer the supernatant into a high-speed round-bottom centrifuge tube while avoid any remaining cell debris

12. Using the swinging-bucket rotor, centrifuge the sample at 21,000×g for 20min at 4°C to pellet the MCPs

13. Remove the supernatant using a serological pipette, taking care not to disturb the MCP pellet

14. Wash the pellet with 2mL Buffer A and 3mL Tris-OTG Buffer and repeat steps 12 and 13

15. Collect the MCP pellet in a 1.5mL microcentrifuge tube and resuspend in 150μL Buffer B; make sure to break up the pellet to separate the MCPs from any remaining cell debris

16. Using the microcentrifuge, centrifuge the sample at 12,000×g for 1min to pellet any remaining cell debris

17. Repeat step 16 two more times, transferring the supernatant into a clean microcentrifuge tube after each spin

18. Store purified MCPs in Buffer B at 4°C, or use immediately

3.3.4. Notes

• If using a fluorescent encapsulation reporter, fluorescence microscopy (see Section 3.1) can be used prior to harvest to check for the formation of Pdu MCPs.

• We use centrifuge rotors that have a capacity for six samples. If processing more than six samples for a specific experiment, the spins can be staggered to allow for alternating centrifugation steps. However, we have found that a maximum of 12 samples should be purified per day, as processing additional samples often leads to timing issues for the centrifugation steps and lower MCP yields.

• A fixed-angle rotor can be used for the 21,000 g spins if it is rated for this speed. However, we find that pelleting the MCPs at a fixed angle results in more broken MCPs in the final sample.

• Purified MCPs are stable at 4°C for approximately 1 week and should be assessed within this time frame. Stability studies have shown that MCPs begin to aggregate after 9 days and begin to deteriorate after 16 days (Kim, Slininger, & Tullman-Ercek, 2014).

• Graf and colleagues recently developed a protocol for smaller scale MCP purifications using only a microcentrifuge (Graf et al., 2017). Depending on the number of samples to be processed and the intended downstream analyses, this MCP “mini-prep” protocol could be used to increase throughput while providing sufficient MCP yields.

• Transmission electron microscopy (TEM) can be used to observe and assess the morphology of purified MCPs (Fig. 1B). We and others have also used thin cell sectioning to observe MCPs within the cell (Fig. 1C) (Cheng, Sinha, Fan, Liu, & Bobik, 2011). Various EM methods are used by the different research groups in the MCP field, and a unified set of best practices specific to MCP imaging has yet to emerge.

3.4. Microcompartment analysis by SDS-PAGE and western blotting

Purified Pdu MCPs can be assessed for composition and encapsulation by SDS-PAGE and western blotting against the cargo of interest (Fig. 7). The total protein concentration of the purified MCP samples should first be measured to normalize loading. These samples can then be separated on appropriate SDS-PAGE gels and the gels subsequently stained by Coomassie or processed for western blotting. Separated Pdu MCPs have a distinctive banding pattern corresponding to the native MCP components, and a band for cargo may also be detected depending on its molecular weight and encapsulation level. Detection of cargo via western blotting for purified MCPs is also indicative of encapsulation.

Fig. 7.

Coomassie-stained gel and αGFP western blot of purified Pdu MCPs. Lanes are labeled with the MCP source strains, which include wild-type S. enterica LT2 (LT2WT), S. enterica LT2 with a knockout of the transcriptional regulator PocR (LT2 ΔpocR), and S. enterica LT2 expressing the encapsulation reporter ssPduD-GFP-ssrA from an arabinose-inducible plasmid (pBAD ssPduD-GFP-ssrA). Note that LT2 ΔpocR acts as a nonassembly control. The Coomassie-stained gel (top) shows the expected banding pattern for LT2 WT and the reporter strain, while no MCP bands are detected for LT2 ΔpocR. Detection of GFP on the western blot (bottom) for the reporter strain indicates encapsulation of ssPduD-GFP-ssrA.

3.4.1. Equipment

• Plate reader

• Heat block

• Standard gel electrophoresis and western blotting equipment (we use the Bio-Rad Mini-PROTEAN and Criterion systems)

• Microwave

• Rocker

• Bio-Rad ChemiDoc XRS+ or other gel imaging system

• Image Lab, ImageJ, or other gel imaging and analysis software

3.4.2. Media, reagents, and consumables

• Bicinchoninic acid (BCA) protein assay kit or Bradford protein assay kit

• 96-well plates

• 4 Laemmli buffer

• 15% SDS-PAGE gels

• Coomassie Stain Solution I: 50% ethanol, 10% acetic acid

• Coomassie Stain Solution II: 5% ethanol, 7.5% acetic acid, Coomassie Brilliant Blue R250

• Polyvinylidene difluoride (PVDF) membrane

• TBST: Tris-buffered saline with 0.05% Tween-20

• Blocking Buffer: 5% nonfat milk (w/v) in TBST

• Appropriate primary antibody and HRP-conjugated secondary antibody using dilutions recommended by the manufacturer for western blotting

• Chemiluminescent substrate (we use Thermo SuperSignal West Pico PLUS)

3.4.3. Procedure

1. Following the manufacturer’s instructions, quantify the total protein concentration for each MCP sample using the BCA or Bradford protein assay kit

2. Denature the MCP samples by boiling in Laemmli buffer at 95°C for 5–10min

3. Load samples onto 15% SDS-PAGE gels, normalized by the measured total protein concentrations

   1. Sample concentrations can be normalized prior to loading by diluting with 1×Laemmli buffer. This allows equal volumes of each sample to be loaded, which results in consistent band widths for densitometry analysis
   2. We typically load between 1.5 and 6μg total protein, depending on the MCP yield and the intended application

4. Run gel at 120V for 90min to separate MCPs

5. Carry out Coomassie staining and/or western blotting per standard procedures:

   1. For Coomassie staining, place gel in Stain Solution I, heat until boiling in microwave, and incubate while rocking until cooled to room temperature. Next, transfer gel to Stain Solution II, heat until boiling in microwave, and incubate while rocking for at least 1h. Finally, transfer the gel to water to destain overnight while rocking; alternatively, a destain solution could be used
   2. For western blotting, transfer proteins to a PVDF membrane and block with 5% milk-TBST for at least 1h. Probe with a primary antibody against the cargo of interest for 2h at room temperature or overnight at4°C.WashthemembranewithTBST(4×5min)andthenprobewith a corresponding HRP-conjugated secondary antibody for 30–60min. Wash the membrane again (4 5min) with TBST prior to imaging
6. Using the ChemiDoc XRS+ or other imaging system:

   1. For Coomassie staining, image the destained gel
   2. For western blotting, develop and image the blot using a chemiluminescent substrate. Follow the manufacturer’s instructions for substrate preparation and application. Image the blot to saturation using signal accumulation mode and select an image just before saturation. In our hands, imaging every 10s for 5min is sufficient to produce the optimal image. If background is too high in signal accumulation mode, manually set the exposure time to produce a single image

7. Carry out densitometry or other analysis using Image Lab or ImageJ software

3.4.4. Notes

• Sample loading can be further normalized based on the densitometry results for specific MCP bands to correct for residual cell debris or other contaminants that may impact the total protein concentrations of the MCP samples.

• We find that 15% SDS-PAGE gels run under the listed conditions are adequate for resolving the bands corresponding to the native Pdu MCP components. However, the molecular weight of the encapsulated cargo should also be considered when selecting an appropriate SDS-PAGE gel. The acrylamide concentration can be adjusted as needed, and gradient or stacked SDS-PAGE gels can also be used. This may require the run voltage and time to be adjusted as well.

• The western blotting protocol can vary significantly depending on the antibodies, imaging substrate, and imaging system used. We suggest using this protocol as a starting point; if other kits and systems are used, follow the manufacturer’s instructions and optimize as necessary.

• Cargo encapsulation levels can be assessed by densitometry of the western blot or the corresponding band on the Coomassie-stained gel. The intensities of the bands correspond to the relative encapsulation levels for each sample.

• When a primary antibody against a cargo of interest is not available, we recommend appending a FLAG-tag to the cargo and probing against this epitope.

3.5. In vitro enzyme assays

The activity of enzymes encapsulated within Pdu MCPs can be measured using in vitro enzyme assays. For these assays, purified MCPs encapsulating the enzyme are incubated with the corresponding substrate, and the sample is monitored for the formation of product. Here, we will discuss the activity assay for encapsulated β-galactosidase. Our lab and others have successfully encapsulated this enzyme within Pdu MCPs and found that it retained its activity (Jakobson et al., 2016; Wagner et al., 2017). We specifically used β-galactosidase tagged with the N-terminal signal sequence ssPduP and C-terminal FLAG- and His-tags. Purified MCPs encapsulating this β-galactosidase fusion protein were then incubated with the substrate o-nitrophenyl-β-galactoside (ONPG), and the activity of β-galactosidase was assessed by monitoring the formation of o-nitrophenol (Fig. 8A and B). The protocol used for this activity assay is provided here, but it should be noted that the overall workflow would be generalizable to other enzymes, substrates, product detection methods, and assay kits (Fig. 8A).

Fig. 8.

Assessing enzymes encapsulated within Pdu MCPs. (A) Diagram of the workflow for encapsulating and assessing enzymes within Pdu MCPs. Enzymes are first tagged with a signal sequence to mediate encapsulation, as well as FLAG- and His-tags for detection and purification. Pdu MCPs coexpressed with these tagged enzymes are then purified and assessed for encapsulation and activity. Encapsulation is confirmed by western blotting against the FLAG epitope, while enzyme activity is assessed by incubating the purified MCPs with the appropriate substrate and monitoring for product formation. For β-galactosidase, the MCPs are incubated with o-nitrophenyl-β-galactoside (ONPG) and the sample is monitored for the formation of o-nitrophenol, which has a yellow color that can be detected by a spectrophotometer at 420nm.(B) Top: Activity of purified MCPs encapsulating a β-galactosidase fusion protein (ssPduP-LacZ-1xFLAG-6xHis), which was expressed in S. enterica LT2 under different MCP induction conditions. Bottom: Western blot of purified MCPs against the FLAG epitope, indicating encapsulation of the β-galactosidase fusion protein. (C) Prior to assessing activity, purified MCPs encapsulating His-tagged enzymes (black circles) can be incubated with nickel-coated magnetic beads to remove copurified, unencapsulated enzymes from the sample. Panel (B): This panel is reproduced from Jakobson, C. M., Chen, Y., Slininger, M. F., Valdivia, E., Kim, E. Y., & Tullman-Ercek, D. (2016). Tuning the catalytic activity of subcellular nanoreactors. Journal of Molecular Biology, 428(15), 2989–2996. https://doi.org/10.1016/j.jmb.2016.07.006. Panel (C): This panel is reproduced from Wagner, H. J., Capitain, C. C., Richter, K., Nessling, M., & Mampel, J. (2017). Engineering bacterial microcompartments with heterologous enzyme cargos. Engineering in Life Sciences, 17(1), 36–46. https://doi.org/10.1002/elsc.201600107.

3.5.1. Equipment

• Plate reader

• Water bath

3.5.2. Media, reagents, and consumables

• BCA protein assay kit or Bradford protein assay kit

• 96-well plates

• Assay Buffer: 50mM potassium phosphate, 300mM sodium chloride, 5mM β-mercaptoethanol, 1mI magnesium sulfate, pH to 8.5 using KOH

• Concentrated Substrate Solution: 10mM ONPG in water

• Buffer B: 50mM Tris–HCl, 50mM KCl, 5mM MgCl₂, 1% 1,2-PD; pH to 8.0 using KOH

3.5.3. Procedure

1. Follow growth and induction steps as outlined in Section 2.2; here, the coexpressed cargo is the β-galactosidase fusion protein (ssPduP-LacZ-1×FLAG-6xHis)

2. Carry out MCP purification as outlined in Section 3.3

3. Following the manufacturer’s instructions, quantify the total protein concentration of the purified MCP sample using the BCA or Bradford protein assay kit

4. Confirm encapsulation of the β-galactosidase fusion protein within the purified MCPs by western blotting against the FLAG epitope as outlined in Section 3.4

5. Prepare assay buffer and concentrated substrate solution and preheat both to 37°C in a water bath

6. In a 96-well plate, combine 2μL of diluted MCP sample, 10μL of concentrated substrate solution, and 88μL of assay buffer for a total assay volume of 100Ml

   1. The added MCP amount should be optimized to ensure that activity measurements are within the linear range of the assay
   2. Note that the final substrate concentration for the assay is 1mM ONPG
7. Measure formation of o-nitrophenol at 420nm using the plate reader

   1. Measurements should be taken every 30s for 10min
   2. The samples should be held at 37°C for the duration of the assay

8. Determine the activity and normalize by the total protein concentration for each sample

3.5.4. Notes

• Wagner and colleagues implemented an additional wash step between MCP purification and the in vitro enzyme assays to remove any copurified, unencapsulated enzymes (Wagner et al., 2017). Specifically, the MCP samples were incubated with nickel-coated magnetic beads that bound any His-tagged enzymes that were not encapsulated by intact MCP shells (Fig. 8C). This step ensured that the measured activities resulted only from encapsulated enzymes.

• Some enzymes and assay kits may not be compatible with the standard buffers and detergents used for MCP purification and storage. In such cases, buffer exchange by dialysis can be used to wash and store MCPs in a more compatible buffer system.

• Not all substrates and products will be able to enter and exit the MCPs through the selective pores of the native Pdu shell proteins. To determine enzyme activity in such cases, MCPs can be broken by dialysis and sonication, and assessed as described (Chowdhury et al., 2015; Fan et al., 2010).

4. Summary and conclusion

Using Pdu MCPs as tunable nanobioreactors would capitalize on powerful strategies to improve heterologous pathway expression and organization. In particular, Pdu MCPs have the potential to enhance metabolic engineering strategies that currently suffer from significant bottlenecks. Toward this end, current efforts seek to use various signal sequences for the encapsulation of heterologous cargo and different techniques to control and assess cargo encapsulation and activity levels. However, as more complex pathways are encapsulated, additional tools will be needed to efficiently tune the levels of multiple enzymes and to control the selective permeability of substrates, cofactors, or products. Moving forward, the use of genomic expression systems will enable controllable cargo encapsulation that is more stable and consistent than current plasmid-based methods. Furthermore, efforts in shell protein engineering will produce MCPs with tunable sizes and diffusive properties, making these scaffolds more widely applicable to diverse and increasingly complex pathways.