The Adaptor Protein p62/SQSTM1 Targets Invading Bacteria to the Autophagy Pathway¹

Immune surveillance for intracellular bacteria in mammalian cells plays a crucial role in cell survival and host defense. Recently, macroautophagy (hereafter referred to as autophagy), an important cellular pathway for the degradation of long-lived proteins and damaged organelles by delivering them to the lysosome (1, 2), was shown to be involved in the innate immune response to bacteria (2, 3). Autophagy is known to target bacteria early after invasion and protect the cytosol from bacterial colonization (4, 5). However, the mechanisms by which bacteria are targeted by autophagy are not known (6).

Salmonella enterica serovar Typhimurium (S. typhimurium)³ is a Gram-negative bacterial pathogen that can infect a variety of hosts (7). Following S. typhimurium invasion, most bacteria reside and replicate within intracellular compartments termed Salmonella-containing vacuoles (SCVs) (8). However, it is now appreciated that a significant fraction (typically 25%) of S. typhimurium is targeted by autophagy, becoming microtubule-associated protein 1 L chain-3 positive (LC3⁺, an autophagy marker) (5). Autophagy was found to restrict the intracellular growth of S. typhimurium (5). Bacteria targeted by autophagy colocalize with polyubiquitinated proteins (5), suggesting ubiquitination of host/bacterial proteins may be involved in autophagy of S. typhimurium.

p62/SQSTM1 (hereafter referred to as p62) is involved in cell signaling, receptor internalization, and protein turnover (9, 10, 11). Recently, p62 was shown to directly bind ubiquitinated proteins (via its C-terminal ubiquitin-associated (UBA) domain) and LC3 (via a newly identified LC3 interacting region (LIR)) (12, 13, 14, 15). It has been suggested that p62 acts as an adaptor by binding ubiquitinated protein aggregates and targeting them for degradation by autophagy, therefore protecting the cytosol from the toxic effects of misfolded or mutated proteins (10, 13, 14). In addition to ubiquitinated protein aggregates, peroxisomes labeled with ubiquitinated proteins were recently found to be degraded by autophagy in a p62-dependent manner (16). Therefore, p62 plays an important role in cellular homeostasis by mediating the autophagy of cargo of various sizes tagged with ubiquitinated proteins. Since S. typhimurium targeted by autophagy is often associated with ubiquitinated proteins (5), we hypothesized that p62 may play an adaptor role in the process of autophagy of bacteria.

Wild-type S. typhimurium SL 1344 and bacteria expressing red fluorescence protein (RFP) (17) were used for these studies. For invasion by S. typhimurium, late-log bacterial cultures were used for infecting cells and prepared via a method optimized for bacterial invasion (18). HeLa human epithelial cells were obtained from the American Type Culture Collection. Wild-type and autophagy-deficient (Atg5^−/−) mouse embryonic fibroblasts (MEFs) have been previously described (19). All cells were cultured in DMEM plus 10% FBS without antibiotics and seeded in 24-well tissue culture plates on glass coverslips 16–24 h before use.

p62 (no. M-010230-00) and siGenome nontargeting siRNA pool no. 2 (no. D-001206-14-20) were from Dharmacon. HeLa cells were transfected with 50 nM each siRNA 48 h before infection using Oligofectamine (Invitrogen) according to the manufacturer’s instruction. Successful p62 knockdown was confirmed by Western blot for each experiment. Plasmids GFP-LC3 (20), GFP-p62-SR (siRNA-resistant p62 tagged with GFP), and GFP-p62-SR-ΔLIR (13), or GFP-p62-SR-ΔUBA was transfected into cells using GeneJuice (Oncogene Research Products) according to the manufacturer’s instructions. To construct the GFP-SR-p62-ΔUBA plasmid pENTR-p62-SR-ΔUBA was first made by exchanging a 1219-bp NheI-EcoRV fragment in pENTR-p62-ΔUBA (10) with a similar fragment from pENTR-p62-SR (13). Finally, GFP-p62-SR-ΔUBA was made by a Gateway LR reaction into pDestEGFP-C1 (21).

Cells were fixed in 2.5% paraformaldehyde in PBS (pH 7.2) for 10 min at 37°C. Fixed cells were stained as previously described (18). Mouse mAb to p62 was from BD Biosciences; rabbit polyclonal Ab to p62 was from Santa Cruz Biotechnology. Rabbit polyclonal Ab to S. typhimurium (Salmonella O anti-serum group B factors 1, 4, 5, and 12) was from Difco Laboratories. FK2 mAb was from BIOMOL. For some experiments, FK2 Ab was conjugated to Alexa Fluor 568 by using the Zenon Alexa Fluor 568 mouse IgG labeling kit (Molecular Probes). Samples were analyzed using a Zeiss Axiovert microscope (×63 objective) and LSM 510 software. Confocal images were imported into Adobe Photoshop and assembled in Adobe Illustrator for labeling. Colocalization quantifications were performed using a Leica DMIRE2 epifluorescence microscope.

For transmission EM, cells were fixed in 2% glutaraldehyde overnight at room temperature. Samples were treated with 1% OsO₄ in 0.1 M Sorenson’s phosphate buffer, stained with 1% aqueous uranyl acetate, dehydrated in a graded series of ethanol, and embedded in Epon. Samples were then sectioned, stained with 2% uranyl acetate, then 0.2% lead citrate, and examined on a FEI Tecnai 20 transmission electron microscope operating at 80 kV.

Cells infected with Salmonella were fixed for immuno-EM in 4% formaldehyde and 0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). Cells were pelleted with 10% gelatin at 37°C, solidified at 4°C, and small blocks were infused with 2.3 M sucrose before mounting on specimen holders and freezing in liquid nitrogen. Sections of 70–90 nm were cut at −110°C on a Leica Ultracut and picked up with a 1:1 mixture of 2.3 M sucrose and 2% methyl cellulose on formvar/carbon-coated grids. Immunolabeling was essentially performed as described earlier (22), using mouse mAbs against p62 (BD Biosciences) and Ab to S. typhimurium (Salmonella O anti-serum group B factors 1, 4, 5, and 12; Difco Laboratories) followed by protein A-gold of 10 nm and 15 nm (CMC). We used a secondary bridging R-anti-M Ab (Dako) as intermediate after the monoclonal p62 Ab. Glutaraldehyde (0.1%) in 0.1 M PBS was used as blocking step in the double labeling procedure. Specimens were observed in a JEOL JEM-1230 at 60–80 kV and images were recorded with a digital camera (Morada), using iTEM software (both from Soft Imaging System). Further image processing was performed using Adobe Photoshop CS2 software.

Colocalization quantifications were performed by direct visualization on a Leica DMIRE2 epifluorescence microscope. At least 100 bacteria were counted for each condition in each experiment. At least three independent experiments were performed. The means ± SD is shown in the figures, and p values were calculated using a two-tailed Student’s t test for Figs. 1,D, 2,C, and 6,B; p values for Figs. 3,D, 4,B, and 5,B were calculated using a one-way ANOVA and Bonferroni posttest; p values for Fig. 7 were calculated using two-way ANOVA and Bonferroni posttests.

To examine the localization of p62 after bacterial invasion, we infected HeLa cells with wild-type S. typhimurium and co-immunostained with an Ab to p62 and the FK2 mAb to ubiquitinated proteins. Since it is known that cytosolic bacteria are targeted by the ubiquitin system and become associated with ubiquitinated proteins, ubiquitination can be used as a marker for bacteria exposed to the cytosol (23). As shown in Fig. 1,A, a population of intracellular S. typhimurium was found to colocalize with p62. A kinetic study showed that colocalization of S. typhimurium with p62 peaked at 1 h postinfection (p.i.) with ∼35% of total intracellular bacteria associated with p62 (Fig. 1,B). The p62 targeting was transient, as colocalization levels dropped to ∼2% at 10 h p.i. (Fig. 1 B). Similar levels of colocalization have been previously observed for a ubiquitinated protein⁺ (Ub⁺) S. typhimurium population (5).

Bacteria that were p62⁺ also labeled strongly for ubiquitinated proteins at 1 h p.i. (Fig. 1,A). We observed that 82% of Ub⁺ S. typhimurium colocalized with p62, vs only ∼5% for the Ub⁻ population of intracellular S. typhimurium (Fig. 1,D). Consistent with this, upon examination of p62⁺ bacteria we observed that ∼96% of these colocalized with ubiquitinated proteins, vs only ∼4% for p62⁻ S. typhimurium (Fig. 1 D). This suggests p62 association with bacteria is an early event and is possibly recruited via binding to ubiquitinated proteins associated with bacteria. In support of this idea, p62 association with bacteria was not observed in cells infected with ΔinvA/inv S. typhimurium (data not shown), a mutant strain of S. typhimurium that does not enter the cytosol after invasion, and is not associated with ubiquitinated proteins (5). Taken together, these data indicate that ubiquitinated protein binding is required for p62 association with bacteria.

Bacteria targeted by autophagy are associated with ubiquitinated proteins (5). Therefore, we examined the association of p62 with the intracellular population of bacteria targeted by autophagy. HeLa cells were transfected with RFP-LC3 and infected with wild-type S. typhimurium. LC3 is conjugated to the autophagosome membrane and is a well-characterized marker for autophagy (2, 20). Approximately 78% of LC3⁺ bacteria colocalized with p62, compared with only ∼5% for LC3⁻ bacteria (Fig. 1, C and D). These data demonstrate that p62 is associated with a population of S. typhimurium targeted by autophagy at an early stage after invasion.

Cells infected with S. typhimurium were analyzed by transmission EM. A population (∼45%) of S. typhimurium was surrounded by multilamellar membranous structures at 1 h p.i. (Fig. 2,A). These multilamellar membranes around the bacteria resembled the autophagic membranes surrounding ΔicsB Shigella flexneri (24). To determine whether the multilamellar structures are formed by autophagy, a MEF cell line lacking expression of the Atg5 gene (Atg5^−/−) was infected with S. typhimurium for 1 h. Atg5 is essential for autophagosome formation; cells lacking Atg5 are completely deficient in autophagy, and no LC3 recruitment to bacteria is observed (5, 19). In Atg5^−/− MEF cells, a significantly lower fraction of bacteria was observed in multilamellar compartments, compared with wild-type MEFs (Fig. 2,C). It is noteworthy that a population of bacteria (∼20%) was observed in multilamellar structures in autophagy-deficient cells, indicating that bacterial virulence factors may promote the formation of these structures independently of host autophagy. Consistent with a drop in multilamellar structures around intracellular bacteria, we observed a higher fraction of bacteria within single membrane SCVs in Atg5^−/− MEFs (Fig. 2 B). Therefore, autophagy of S. typhimurium promotes the formation of multilamellar structures surrounding bacteria.

Next, we examined the localization of p62 in HeLa cells infected with S. typhimurium for 1 h (the time of peak autophagy and p62 localization with bacteria) by immunogold EM. As shown in Fig. 2,D, p62 was observed at the multilamellar autophagy membrane surrounding S. typhimurium. Double-immunogold EM labeling of both S. typhimurium (15 nm gold particle) and p62 (10 nm gold particle) confirmed that p62 in most cases localized between the bacterium and multilamellar membranes; however, in some instances p62 also seemed to associate with the bacterial surface (Fig. 2 E). Therefore, p62 is associated with bacteria targeted by autophagy.

To investigate whether p62 plays a role in the autophagy of bacteria, p62 expression was targeted by siRNA, and complemented with siRNA-resistant (SR) p62 mutant constructs (Fig. 3,A). Additionally, the cells were transfected with RFP-LC3 and infected by wild-type S. typhimurium. The knockdown and construct expression were confirmed by Western blotting (Fig. 3,B). Knockdown of endogenous p62 reduced the levels of S. typhimurium targeted by autophagy by ∼50% (Fig. 3, C and D). Therefore, p62 plays a significant role in autophagy of bacteria. As predicted, cells with p62 knockdown, and then complemented with GFP-p62-SR, showed normal autophagy of bacteria (Fig. 3, C and D). However, autophagy was not restored in p62 siRNA-treated cells expressing either GFP-p62-SR-ΔLIR (lacking LIR domain) or GFP-p62-SR-ΔUBA (lacking UBA domain). Therefore, p62 requires both its UBA and LIR domains for its role in autophagy of S. typhimurium, consistent with its function as an adaptor protein.

Using this complementation strategy we examined the mechanism of targeting of p62 to bacteria. First, we focused on the recruitment of p62 to Ub⁺ bacteria. As expected, GFP-p62-SR and GFP-p62-SR-ΔLIR were recruited to Ub⁺ S. typhimurium in p62 siRNA-treated cells (Fig. 4), with levels of colocalization similar to endogenous p62 (Fig. 1,D). However, cells complemented with GFP-p62-SR-ΔUBA showed a significant decrease in the levels of colocalization, indicating that the UBA domain is critical for the recruitment of p62 to Ub⁺ S. typhimurium (Fig. 4).

Next we examined recruitment of p62 to LC3⁺ bacteria. We observed that GFP-p62-SR was recruited to LC3⁺ S. typhimurium in p62 siRNA-treated cells (Fig. 3,C), with levels of colocalization similar to endogenous p62 (Fig. 5,B). This is consistent with normal levels of autophagy observed when p62 expression is complemented by the GFP-p62-SR construct (Fig. 3,D). We also examined recruitment of GFP-p62-SR-ΔLIR and GFP-p62-SR-ΔUBA to LC3⁺ bacteria in p62 siRNA-treated cells. It must be borne in mind that in these cells the levels of autophagy is ∼50% inhibited (Fig. 3,D). The fact that autophagy is not abolished by p62 knockdown suggests the existence of an alternate pathway for regulating autophagy of bacteria. Interestingly, we observed recruitment of GFP-p62-SR-ΔUBA to LC3⁺ bacteria in p62 siRNA-treated cells (Fig. 5). In contrast, GFP-p62-SR-ΔLIR was not recruited to LC3⁺ bacteria in p62 siRNA-treated cells (Fig. 5). Therefore, p62 is recruited to Ub⁺ bacteria primarily via its UBA domain. However, p62 lacking the UBA domain can be recruited to autophagosomes containing LC3⁺ bacteria via its LIR domain. These observations provide a basis for understanding the association of p62 with both the bacterial surface and multilamellar autophagic membranes surrounding bacteria (Fig. 2 D) and are consistent with the finding that p62 is recruited to autophagosomes via its LIR domain during starvation-induced autophagy (13, 15).

To further examine the role of LC3 binding in the recruitment of p62 to bacteria, we utilized Atg5^−/− MEFs. Both wild-type and Atg5^−/− MEFs were infected with wild-type S. typhimurium expressing RFP and immunostained for p62. Intracellular bacteria (∼20%) colocalized with p62 at 1 h p.i. in wild-type and Atg5^−/− MEFs (Fig. 6). At 4 h p.i., the colocalization of p62 with S. typhimurium in wild-type MEF dropped to ∼10% (Fig. 6,B), consistent with our observations in HeLa cells (Fig. 1,B). In contrast, S. typhimurium colocalization with p62 was sustained in Atg5^−/− MEF (Fig. 6 B). This is consistent with an increase in the number of cytosolic bacteria associated with ubiquitinated proteins observed in these autophagy-deficient MEFs (5). Taken together, these findings demonstrate that p62 association with S. typhimurium can occur independently of autophagy.

Autophagy restricts the intracellular growth of bacteria (2, 5). To determine the importance of p62 in autophagy of bacteria, we assessed bacterial replication in cells where p62 expression was targeted by siRNA. We enumerated the number of intracellular S. typhimurium over a 10-h infection period using immunofluorescence and categorized as the percentage of infected cells containing a given number of intracellular bacteria (Fig. 7). Intracellular bacterial growth over the 10-h infection period was indicated by a decrease in the percentage of cells with lower numbers of bacteria and an increase in the percentage of cells with higher numbers of bacteria. At early time points, the number of bacteria per cell was similar between control siRNA and p62 siRNA-treated cells (Fig. 7). However, at 10 h p.i., p62 siRNA-transfected cells showed a significantly greater percentage of cells with high numbers of intracellular bacteria than control siRNA-treated cells (∼57% vs ∼30%, respectively, for >20 bacteria/cell), as well as a significantly smaller percentage of cells with few numbers of intracellular bacteria (∼14% vs 28%, respectively, for 1–5 bacteria/cell). Additionally, a time course analysis revealed that the effect of p62 siRNA was similar to atg12 siRNA treatment (supplemental Fig. 2).⁴ These findings demonstrate that p62 expression inhibits intracellular bacterial replication, consistent with its role in autophagy of bacteria.

In previous studies, p62 has been found in many cellular inclusions, including Mallory bodies in chronic liver disease (25), aggresome-like induced structures induced by puromycin treatment (13), Lewy bodies in Parkinson’s disease, and neurofibrillary tangles in Alzheimer’s disease (25, 26). It has been suggested that p62 plays a cytoprotective role by providing a scaffold for misfolded or mutated proteins, initiating formation of aggregates and then targeting these for autophagic degradation (10) (Fig. 8, aggrephagy). Additionally, recent studies suggest that p62 might contribute to midbody ring disposal by autophagy at the end of cytokinesis (27). Our studies reveal that p62 can also target intracellular bacteria to the autophagy pathway.

As shown here, the recruitment of p62 is associated with ubiquitinated proteins localized to the bacteria. The identity of the host/bacterial ubiquitinated proteins is currently unknown and will be an important subject for future studies. It is noteworthy that p62 preferentially binds lysine 63-linked polyubiquitinated proteins (28, 29, 30). We demonstrate that p62 is required for autophagy of intracellular bacteria, but p62 recruitment is independent of autophagy. We propose a model whereby bacteria are decorated first with ubiquitinated proteins, followed by recruitment of p62, and are subsequently targeted by autophagy and delivered to lysosomes for degradation (Fig. 8, xenophagy). Interestingly, recently published data suggest p62 plays a similar role in autophagy of peroxisomes (16) (Fig. 8, pexophagy). Our study suggests that p62 plays an adaptor role to mediate the specific targeting of S. typhimurium by autophagy, thereby restricting the intracellular growth of these bacteria. This is the first study demonstrating that p62 plays a role in host innate immune recognition of intracellular bacteria. Whether p62 also plays a role in autophagy of other intracellular bacteria is an important question for future studies. A recent study shows that NBR1 (neighbor of BRCA1 gene 1), which has a very similar domain structure to p62, acts as a receptor for selective autophagosomal degradation of ubiquitinated protein aggregates (31). To investigate if NBR1 is important for autophagy targeting of bacteria, we used siRNA to knockdown endogenous NBR1 and studied its effect in recruitment of LC3 to S. typhimurium. Our data suggested that NBR1 does not play any major role in the autophagy targeting of bacteria (supplemental Fig. 1). In summary, p62 appears to be an important adaptor protein, or cargo receptor, that mediates surveillance of multiple cargoes marked by ubiquitinated proteins, including protein aggregates, organelles, and bacteria, to regulate the role autophagy plays in cellular homeostasis and immunity.

While this manuscript was in press, two studies were published that describe a similar role for p62 in xenophagy. Dupont et al. demonstrated that p62 mediates xenophagy of Shigella flexneri (2009, Cell Host Microbe, 6: 137–149). In a study of Listeria monocytogenes, Yoshikawa et al. demonstrated that a non-motile actA mutant is targeted by xenophagy through p62 recruitment (2009, Nat. Cell Biol., advanced online publication). These studies, and our own, establish an important role for p62 in the regulation of autophagy as an innate immune defence.

We thank Drs. Nat Brown, Nicola Jones, Noboru Mizushima, and Tamotsu Yoshimori for providing reagents. We thank M. Woodside and P. Paroutis for assistance with confocal microscopy. We also thank Drs. Jorge Moscat and Gabriel Núñez for experimental advice and members of the Brumell Laboratory for constructive comments on the manuscript.

The authors have no financial conflicts of interest.