The PduM Protein Is a Structural Component of the Microcompartments Involved in Coenzyme B₁₂-Dependent 1,2-Propanediol Degradation by Salmonella enterica

Abstract

Diverse bacteria use proteinaceous microcompartments (MCPs) to optimize metabolic pathways that have toxic or volatile intermediates. MCPs consist of metabolic enzymes encased within a protein shell that provides a defined environment. In Salmonella enterica, a MCP is involved in B₁₂-dependent 1,2-propanediol utilization (Pdu MCP). In this report, we show that the protein PduM is required for the assembly and function of the Pdu MCP. The results of tandem mass spectrometry and Western blot analyses show that PduM is a component of the Pdu MCP. Electron microscopy shows that a pduM deletion mutant forms MCPs with abnormal morphology. Growth tests and metabolite measurements establish that a pduM deletion mutant is unable to form functional MCPs. PduM is unrelated in sequence to proteins of known function and hence may represent a new class of MCP structural proteins. We also report a modified protocol for the purification of Pdu MCP from Salmonella which allows isolation of milligram amounts of MCPs in about 4 h. We believe that this protocol can be extended or modified for the purification of MCPs from diverse bacteria.

INTRODUCTION

Bacterial microcompartments (BMCs) are a diverse family of proteinaceous organelles used to optimize metabolic pathways that have toxic or volatile intermediates (7, 12, 15, 62). They are polyhedral in shape and 100 to 150 nm in cross-section and consist of a protein shell that encapsulates specific metabolic enzymes (29). Sequence analyses indicate that proteinaceous microcompartments (MCPs) are produced by 20% to 25% of bacteria and function in 7 or more different metabolic processes (3, 7, 15). MCPs play a major role in global carbon fixation (2, 12, 33), are linked to bacterial pathogenesis (11, 17, 27, 32, 57, 60), and have a number of potential biotechnology applications, since their protein shells provide a basis for engineering synthetic protein cages for use as drug delivery vehicles or as nanoscale intracellular chemical reactors for the production of chemicals (18, 22, 42, 43, 61, 63).

BMCs share related protein shells but differ in enzyme content according to their function (7, 12, 15, 62). The carboxysome MCP, which is part of a CO₂-concentrating mechanism used to enhance autotrophic CO₂ fixation, encapsulates RuBisCO and carbonic anhydrase (12, 62). Other MCPs encase 1,2-propanediol (1,2-PD) or ethanolamine-degradative enzymes or proteins of unknown function (7, 12, 15, 62). The function of MCP shells is to restrict the outward diffusion of a toxic or volatile metabolic intermediate and channel it to downstream enzymes. In the case of the carboxysome, the shell restricts the outward diffusion of CO₂ produced by carbonic anhydrase, which elevates the CO₂ concentration in the vicinity of RuBisCO, enhancing carbon fixation (21, 46). The protein shell of the MCPs used for 1,2-PD and ethanolamine degradation restricts the diffusion of propionaldehyde and acetaldehyde, respectively, to help prevent toxicity and/or carbon loss (26, 45, 48, 49, 54). The shells of bacterial MCPs are usually made from 5 to 10 different proteins, most of which have bacterial microcompartment (BMC) domains (19, 28, 34, 35, 50, 55, 56). Crystallographic studies have shown that BMC domain proteins have central pores that differ in charge and size (19, 28, 34, 35, 50, 55, 56). These pores are proposed to control the inward movement of substrates, cofactors required by the encapsulated enzymes, as well as the outward movement of reaction products (19, 28, 34, 35, 50, 55, 56).

Prior work by our laboratory showed that an MCP is used for coenzyme B₁₂-dependent 1,2-PD utilization (Pdu MCP) by Salmonella enterica (7, 9, 10, 19, 25, 26, 36, 37, 52). 1,2-PD is a major product of the anaerobic degradation of the common plant sugars rhamnose and fucose, and it is thought to be an important carbon and energy source in anoxic environments such as the large intestines of higher animals (40). In addition, several independent studies have suggested that the ability to degrade 1,2-PD contributes to pathogenesis by Salmonella and Listeria (11, 17, 27, 32, 57, 60). The MCP used for 1,2-PD consists of a protein shell that encapsulates 1,2-PD degradative enzymes (25, 26). Mutants unable to properly assemble the Pdu MCP accumulate propionaldehyde to levels that cause DNA damage and cell toxicity (25, 26, 52). The Pdu MCP is thought to mitigate toxicity by confining propionaldehyde and channeling it to downstream enzymes (9, 25, 26, 52). The enzymes and proteins used for 1,2-PD degradation and MCP assembly are encoded by a single cluster of contiguous genes: pocR, pduF, pduABB′CDEGHJKLMNOPQSTUVWX (8–10, 13, 23, 24, 31, 36, 37, 47). PduCDE, PduL, PduP, PduQ, and PduW are enzymes for 1,2-PD catabolism (Fig. 1). This process begins with the conversion of 1,2-PD to propionaldehyde by coenzyme B₁₂-dependent diol dehydratase (PduCDE). Propionaldehyde is then converted to propionate by coenzyme A (CoA)-dependent propionaldehyde dehydrogenase (PduP), phosphotransacylase (PduL), and propionate kinase (PduW) (9, 41) or to 1-propanol by propionaldehyde dehydrogenase (PduQ). The PduGH, PduO, and PduS proteins are used to provide diol dehydratase (PduCDE) with its required cofactor, coenzyme B₁₂ (14, 31, 39). Overall, 1,2-PD degradation provides a source of ATP and propionyl-CoA which feeds into central metabolism via the methylcitrate pathway (41). The shell of the Pdu MCP, which is composed of the PduABB′JKNTU proteins, acts as a diffusion barrier that channels propionaldehyde to the PduP enzyme (25, 26, 36, 52, 54). The PduV protein is thought to have a role in spatial orientation of the Pdu MCP within the cytoplasm of the cell (44). The PduM protein is of unknown function, lacks an identifiable BMC domain, and is unrelated in sequence to proteins of known function. In this report, we extend our studies of the Pdu MCP by showing that PduM is a newly identified type of structural protein that is essential for its assembly and function.

Fig 1.

Model for Pdu microcompartment. The dashed line indicates the shell of the MCP, which is composed nine different polypeptides (PduABB′JKMNTU). The first two steps of 1,2-PD degradation are thought to occur in the lumen of the MCP. Its proposed functions are to confine propionaldehyde and channel it to the downstream enzyme in order to minimize toxicity. Abbreviations: 1,2-PD, 1,2-propanediol; Ado-B₁₂, coenzyme B₁₂; PduCDE, coenzyme B₁₂-dependent diol dehydratase; PduP, propionaldehyde dehydrogenase; PduGH, reactivase; PduO, adenosyltransferase; PduS, cobalamin reductase; PduL, phosphotransacylase; PduW, propionate kinase; PduQ, 1-propanol dehydrogenase.

MATERIALS AND METHODS

Bacterial strains, media, and growth conditions.

The bacterial strains used in this study were Salmonella enterica serovar Typhimurium LT2 and derivatives. The rich medium used was lysogeny broth (LB), also known as Luria-Bertani/Lennox medium (Difco, Detroit, MI) (5, 6, 38). The minimal medium used was no-carbon-E (NCE) (4, 58).

Chemicals and reagents.

Antibiotics were from Sigma Chemical Company (St. Louis, MO). Bacterial Protein Extraction Reagent II (BPER-II) was from Pierce (Rockford, IL). Electrophoresis supplies were from Bio-Rad (Hercules, CA). IPTG (isopropyl-β-d-thiogalactopyranoside) was from Diagnostic Chemicals Limited (Charlottesville, Prince Edwards Island, Canada). DNase I (DNase) and 4-(2-aminoethyl)-benzenesulfonyl fluoride HCl (AEBSF) were from Sigma. Restriction enzymes and T4 DNA ligase were from New England BioLabs (Beverly, MA). Other chemicals were from Fisher Scientific (Pittsburgh, PA).

Molecular and genetic methods.

Agarose gel electrophoresis, plasmid purification, PCR, restriction digests, ligation reactions, and electroporation were carried out using protocols previously described (37, 51). Plasmid DNA was purified using Qiagen products (Qiagen, Chatsworth, CA) according to the manufacturer's instructions. Following restriction digestion or PCR amplification, DNA was purified using Promega Wizard PCR Preps (Madison, WI). Restriction digests were carried out using standard protocols (51), and T4 DNA ligase was used according to the manufacturer's directions (New England BioLabs). Electroporation was carried out as described previously (9). P22 transduction was performed as described using P22 HT105/1 int-210 (53). Transductants were tested for phage contamination and sensitivity by streaking on green plates against P22 H5. For complementation studies, the coding sequences of pduM were cloned into pLAC22 via PCR with template pEM55 as previously described (9). Vector pLAC22 allows tight regulation of protein production by IPTG (59). The DNA sequence was verified. The pduM deletion was constructed by the method of Miller and Mekalanos, with modifications previously described (30). PCR was used to verify all deletions as described previously (20).

DNA sequencing and analysis.

DNA sequencing was carried out at the Iowa State University DNA Facility using Applied Biosystems Inc. automated sequencing equipment. The template for DNA sequencing was plasmid DNA purified using Qiagen 100 tips or Qiagen MiniPrep kits. BLAST software was used for sequence similarity searching (1).

Purification of Pdu MCPs.

Growth medium (1× NCE medium, 1 mM MgSO₄, 0.5% succinate, and 0.6% 1,2-PD) (400 ml) was inoculated with 2 ml of an overnight LB culture and incubated at 37°C with shaking at 275 rpm to an optical density at 600 nm (OD₆₀₀) of 1.0 to 1.2. Cells were harvested by centrifugation and washed twice with 40 ml of buffer A (50 mM Tris-HCl [pH 8.0], 500 mM KCl, 12.5 mM MgCl₂, 1.5% 1,2-PD). Cells (∼1 g wet weight) were resuspended in a mixture of 10 ml of buffer A and 15 ml of BPER-II supplemented with 5 mM β-mercaptoethanol, 0.4 mM protease inhibitor AEBSF, 25 mg of lysozyme, and 2 mg of DNase I. The suspension was incubated at room temperature with 60 rpm of shaking for 30 min and then on ice for 5 min. Cell debris was removed by centrifugation twice at 12,000 × g for 5 min at 4°C, and the MCPs were spun down at 20,000 × g for 20 min at 4°C. The pellet was washed once with a mixture of 4 ml of buffer A and 6 ml of BPER-II containing 0.4 mM AEBSF and then was resuspended in 0.5 ml of buffer B (50 mM Tris-HCl [pH 8.0], 50 mM KCl, 5 mM MgCl₂, 1% 1,2-PD) containing 0.4 mM AEBSF. Remaining cell debris was removed by centrifugation at 12,000 × g for 1 min at 4°C performed three times. Purified MCPs were stored at 4°C until used.

SDS-PAGE and Western blot analyses.

Protein concentration was determined using Bio-Rad (Hercules, CA) protein assay reagent with bovine serum albumin (BSA) as a standard. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed using Bio-Rad 10% to 20% gradient Tris-HCl ready gels. Western blot analyses were carried out as previously described (14) with commercially prepared primary rabbit antisera against synthetic peptide IVSRLHRRAQSTAT, which is present at positions 13 to 26 in PduM (Genscript, Piscataway, NJ).

DDH activity assay.

Diol-dehydratase (DDH) assays were performed using the coupled NADH-dependent alcohol dehydrogenase reaction as previously described (10) or the 3-methyl-2-benzothiazolinone hydrazone method as described by Havemann and Bobik (25). One unit of activity is defined as corresponding to the amount of enzyme that catalyzes the formation of 1 μmol of propionaldehyde per min per mg of protein.

Growth studies.

Growth rates were determined using a Synergy HT microplate reader (BioTek, Winooski, VT) as previously described (37). Doubling times were calculated from semi-log plots where doubling time = 0.693/(2.303) (slope of the linear region of the plot).

Propionaldehyde determination.

Cultures (50 ml) were grown in 250-ml Erlenmeyer flasks at 37°C with shaking at 275 rpm in an Innova I2400 incubator shaker (New Brunswick Scientific) as described previously (26). Samples were taken at timed intervals. Cells were removed by centrifugation followed by filtration with Millex-GV syringe filters (Millipore Corporation) (0.22 μm pore size). Propionaldehyde levels were then determined by high-pressure liquid chromatography (HPLC) and by the 3-methyl-2-benzothiazolone hydrazone (MBTH) assay as described previously (52).

EM.

For electron microscopy (EM), strains were grown in 125-ml Erlenmeyer flasks containing 10 ml of NCE minimal medium supplemented with 1 mM MgSO₄, 0.5% succinate, 0.4% 1,2-PD, and 50 μM ferric citrate. Cultures were inoculated at an OD₆₀₀ of 0.1 with an LB overnight culture that had been centrifuged and resuspended in NCE medium with 100 μM MgSO₄. At an optical density at 600 nm of between 1 and 1.2, cells were harvested by centrifugation. Imbedding, sectioning, and electron microscopy were carried out as described previously (9, 26). For the purified MCPs, following the purification, the MCPs were suspended in buffer B at a concentration of 0.1 mg/ml. A 10-μl volume of this solution was placed on a Formvar-coated copper grid. After 1 min of incubation, the excess solution was removed using Whatman 40 (GE Healthcare, Piscataway, NJ) filter paper, and then 10 μl of 2% uranyl acetate (electron microscopy grade) in MilliQ water (Millipore, Billerica, MA) containing 0.1% 1,2-PD was added to the grid. Excess uranyl acetate was removed using Whatman 40 filter paper, and the grids were dried and stored for visualization under a JEOL 2100 scanning transmission EM (JEOL USA, Inc., Peabody, MA).

Matrix-assisted laser desorption ionization–time of flight tandem mass spectrometry (MALDI-TOF MS-MS).

Tandem mass spectroscopy was performed as previously described (14).

RESULTS

Improved protocol for the purification of Pdu MCPs.

Purification of the Pdu MCPs with the modified protocol described here yielded ∼1 mg of MCPs from 400 ml of initial culture. The protocol takes about 4 h to complete, whereas the prior method required about 18 h (25). Electron microscopy of the negatively stained samples obtained by the new protocol showed homogenous, intact MCPs and little cell debris (Fig. 2). The specific activity of the MCP-associated DDH enzyme was 29.8 μmol min⁻¹ mg⁻¹, which compares favorably to the prior method, where the DDH activity of purified MCPs was 27.5 μmol min⁻¹ mg⁻¹ using the same assay method (25). The MCPs prepared this way were structurally stable and active for 24 h as examined by both electron microscopy and DDH assay. This allows sufficient time for the characterization of the microcompartments and their mutant derivatives.

Fig 2.

Electron microscopy of Pdu MCPs purified by an improved protocol reported here.

Sequence analysis and distribution of PduM.

The PduM protein of S. enterica (GI 16765378) is 163 amino acids in length. It is unrelated to proteins of known function and lacks an identifiable BMC domain or other obvious motifs. PduM has over 200 homologues in GenBank (blastp expect value of ≤5 × 10⁻⁴) that are distributed among the genera Salmonella, Shigella, Citrobacter, Listeria, Yersinia, Klebsiella, Escherichia, Tolumonas, and Shewanella. In each of these genera, PduM homologues are encoded by operons predicted to be involved in 1,2-PD degradation, with the exception of Shewanella, where PduM homologues are found in operons that encode a putative MCP of unknown function as well as enzymes with homology to pyruvate formate lyase.

Electron microscopy of a pduM deletion mutant.

A precise deletion of the pduM gene was constructed using a PCR-based method, and its effects on MCP formation were examined by electron microscopy. Many cells and numerous thin sections were examined. Representative images are shown (Fig. 3). The pduM deletion (BE189) formed various cytoplasmic structures when grown on 1,2-PD minimal medium (Fig. 3). Some of these structures were similar in appearance to normal MCPs, but others were large amorphous aggregates that were likely composed of MCP components. About 73% of ΔpduM cells contained one or more abnormal structures that were markedly different from those of normal MCPs, as exemplified in Fig. 3. In contrast, when wild-type Salmonella was grown under similar conditions, >90% of cells contained normal-appearing MCPs and the remaining cells lacked MCPs. Thus, electron microscopy studies showed that the Salmonella pduM deletion mutant was impaired for MCP assembly.

Fig 3.

A pduM deletion mutant forms abnormal MCPs. Left panel: a representative image of wild-type Salmonella enterica serovar Typhimurium LT2. Right panel: a representative image of a BE189 pduM deletion mutant.

Growth phenotypes of a pduM deletion mutant.

Prior studies identified two phenotypes that are characteristic of mutations that disrupt the shell of the Pdu MCP. The first is a phenotype whereby propionaldehyde accumulated to toxic levels and induced a period of growth arrest that lasted from 12 to 20 h during growth on 1,2-PD at saturating B₁₂ levels (100 nM) (26). The second is a phenotype whereby MCP mutants grow substantially faster than the wild type at limiting B₁₂ levels (20 to 50 nM) (16). Here, a pduM deletion mutant was found to have both growth phenotypes characteristic of MCP disruption. At saturating B₁₂ levels, the wild type grew normally and accumulated propionaldehyde to about 2 mM. Under similar conditions, a pduM mutant accumulated propionaldehyde to about 15 mM and underwent a 16-h period of growth arrest (Fig. 4). The pduM deletion mutant was also found to grow faster than the wild type at a limiting B₁₂ level (20 nM). Wild-type Salmonella grew with a doubling time of 15.2 ± 1.3 h with limiting B₁₂, whereas the pduM deletion had a doubling time of 4.9 ± 0.2 h under similar conditions. Thus, the growth phenotypes of the pduM deletion mutant indicated that the shell of the MCP was disrupted, which was consistent with the electron microscopy results.

Fig 4.

Growth phenotypes of a pduM deletion mutant. (A) Growth and propionaldehyde production by wild-type Salmonella (filled and open circles, respectively) and a pduM deletion mutant (filled and open diamonds, respectively) during growth on 1,2-PD with saturating vitamin B₁₂ (100 nM). (B) Growth of wild-type strain (closed circles) and a pduM deletion mutant (open squares) on 1,2-PD with limiting (20 nM) vitamin B₁₂.

Complementation of the growth phenotypes of the pduM deletion.

Studies were done to determine whether the pduM deletion mutant could be complemented by pduM expressed from a plasmid. The vector used for complementation was pLAC22, which allows tight regulation of gene expression by IPTG (59). The results presented in Fig. 5 show that the fast-growth phenotype was complemented when 0.005 mM IPTG was used for inducing pduM expression. This indicated that the observed growth phenotypes were the result of the mutation under investigation and not due to polarity or an unknown mutation. In these studies, we also found that complementation of the pduM deletion was dependent on the level of pduM expression. At IPTG concentrations of 0.002 to 0.007 mM, partial to full complementation was observed. At 0.01 mM IPTG or higher, no complementation was seen. In addition, overexpression of PduM in wild-type LT2 also resulted in fast growth at a limiting (20 nM) vitamin B₁₂ level (Fig. 5), which is an indication of disrupted or broken MCPs. Thus, results indicated that proper MCP formation requires a measured amount of PduM and that larger amounts interfere with proper MCP formation. This is in contrast to the results of our prior studies where overexpression of pduBB′, pduJ, pduK, pduN, pduT, pduU, pduO, pduP, and pduS did not impair MCP formation (14, 16, 31, 36).

Fig 5.

Complementation of a pduM deletion mutant requires a measured amount of PduM. (A) Complementation. The ΔpduM mutant grows faster than the wild type on 1,2-PD at a limiting level (20 nM) of B₁₂, which is typical of mutations that impair MCP formation. This phenotype was corrected by expression of PduM from plasmid pLAC22 when 0.02 mM IPTG was used to induce expression. Filled circles, wild-type Salmonella/pLAC22; filled triangles, ΔpduM/pLAC22; open squares, ΔpduM/pLAC22-pduM. Higher levels of PduM expression did not allow complementation (not shown). (B) Overexpression of pduM from pLAC22 in the wild type resulted in a phenotype indicative of MCP impairment during growth on 1,2-PD at limiting (20 nM) B₁₂. Production of PduM was induced by IPTG. Closed circles, no IPTG; open squares, 0.01 mM IPTG; filled triangles, 0.02 mM IPTG; filled diamonds, 0.1 mM IPTG.

PduM is a component of Pdu MCP.

To examine the cellular location of PduM, MCPs purified from wild-type S. enterica and a pduM deletion mutant (BE189) were analyzed by SDS-PAGE, MALDI-TOF MS-MS, and Western blotting (Fig. 6). A light band near the expected molecular mass of PduM (17.9 kDa) was observed following SDS-PAGE of wild-type MCPs but was missing in the pduM deletion mutant MCPs. Digestion of this band with trypsin followed by MALDI-TOF MS-MS identified six peptide sequences (R·IVEEIVSR·L [944.55 Da], K·IHDALAFGIR·V [1,112.62 Da], R·KIHDALAFGIR·V [1,240.73 Da], R·AQSTATLSVTQLR·D [1,375.76 Da], R·RAQSTATLSVTQLR·D [1,531.85 Da], and R·DADCPALFCQHASLR·I [1,760.82 Da]) that matched that of the PduM protein. This indicated that PduM is a component of the Pdu MCP. PduM localization was further analyzed by Western blotting of crude cell extracts and purified MCPs from wild-type Salmonella and a pduM deletion mutant (Fig. 6B). A band near 17.9 kDa was detected in MCPs purified from wild-type S. enterica, but no PduM band was detected in MCPs purified from the pduM deletion mutant. Moreover, when similar amounts of protein were analyzed, Western blotting readily detected PduM in purified MCPs but not in crude cell extracts (Fig. 6B). This indicated that purified MCPs were enriched in PduM. Cumulatively, the results presented above show that PduM is a component of Pdu MCPs.

Fig 6.

PduM is a component of the Pdu MCP. (A) A 10% to 20% SDS-PAGE gel was stained with Coomassie brilliant blue. Lane 1, molecular mass markers; lane 2, 10 μg of Pdu MCPs purified from wild-type Salmonella; lane 3, 10 μg of Pdu MCPs purified from a pduM deletion mutant. PduM was identified by MS-MS. (B) Western blot for PduM. Lanes 1 and 2, 10 μg of whole-cell extract and 10 μg of purified MCPs from the wild type; lanes 3 and 4, 10 μg of whole-cell extract and purified MCPs from BE189 (ΔpduM).

Our results also indicated that the ΔpduM MCPs were unstable compared to wild-type MCPs. The yield of ΔpduM MCPs (based on protein assay data) was about 14% of the yield of the wild-type MCPs. This was likely due to breakdown of ΔpduM MCPs during purification, which is consistent with their highly aberrant structure (Fig. 3). Studies also showed that the specific activity of DDH in MCPs purified from the pduM mutant was about 42% of the wild-type value. The SDS-PAGE results indicated that this was partly due to the lower purity obtained for ΔpduM MCPs compared to wild-type MCPs (Fig. 6). However, the increased contamination observed by SDS-PAGE appears to have been insufficient to fully account for the measured decrease in DDH activity (Fig. 6). We think the most likely explanation is that DDH was lost from broken MCPs during purification from ΔpduM mutants. Thus, purification studies indicate that the ΔpduM MCPs were unstable, which further supports a structural role for the PduM protein.

DISCUSSION

In this report, we have presented a simplified method that allows isolation of milligram amounts of MCPs in about 4 h whereas our prior protocol described in reference 25 required about 18 h. This protocol should substantially facilitate further biochemical and compositional studies of the Pdu MCP and might be extended to the purification of microcompartments from other bacterial strains. Although methods for the purification of carboxysomes from Halothiobacillus are available, the purification of carboxysomes from cyanobacteria has proven difficult (12). We believe that the purification protocol described here may be extended to the purification of microcompartments from these strains.

We also conducted a series of studies that showed that the PduM protein is an essential structural component of the Pdu MCP that is required in precise amounts. MS-MS and Western blot analyses established that the PduM protein is a component of purified Pdu MCPs. Electron microscopy and growth tests showed that a pduM deletion mutant was impaired for MCP assembly and function. In addition, growth tests showed that a pduM deletion mutant metabolized 1,2-PD at rates similar to (at saturating B₁₂ concentrations) or faster than (at limiting B₁₂ concentrations) that of the wild type, suggesting that it has no direct role in catalysis. The studies presented here also showed that excess PduM impaired MCP formation, since the overproduction, even in the wild-type strain, impaired MCP assembly. This suggests that excess PduM interacts aberrantly with other proteins essential for MCP formation. We also note that prior proteomics analyses failed to identify PduM as an MCP component (25). This was likely due to its low abundance or its relatively high pI (9.56), which could have resulted in migration to the edge of or off the gel during the isoelectric focusing step of the two-dimensional (2D) electrophoresis used in prior work (25). Overall, the results presented in this report show that PduM is an essential structural component of the Pdu MCP that is required in measured amounts. Furthermore, because PduM is unrelated in sequence to other MCP shell proteins, it likely represents a new class of MCP structural protein.

Although it is clear that the PduM protein is essential for proper MCP assembly, its specific function is uncertain. Given that the Pdu MCPs deviate substantially from a regular icosahedral shape, it is possible that they might have two or more types of vertex protein (25). Prior studies indicated that carboxysome vertex proteins are much less abundant than the proteins that make up the flat surfaces of the shell (55). SDS-PAGE analyses of purified Pdu MCPs showed that PduM has low relative abundance compared to other Pdu shell proteins (Fig. 6). We also found that a pduM deletion mutant formed MCPs that are mostly abnormally shaped, which is consistent with a pivotal structural role such as that of a vertex protein. Prior studies on the carboxysome indicated that CcmL and homologues (PduN) are vertex proteins (55). Thus, we speculate that, given its more irregular structure, two types of vertex proteins (PduM and PduN) are required for assembly of the Pdu MCP.

As mentioned above, bacterial MCPs provide a basis for engineering protein cages for use as intracellular bioreactors for the improved production of chemicals or for use as drug delivery vehicles. A detailed understanding of the structure and function of MCPs would help realization of this goal. In this report, we have showed that PduM represents a new class of MCP structural protein that is required for MCP assembly. We think that this will be helpful for raising our understanding of the Pdu MCP to the level needed for biotechnology applications.