Endophilin B2 facilitates endosome maturation in response to growth factor stimulation, autophagy induction, and influenza A virus infection

Endocytosis, and the subsequent trafficking of endosomes, requires dynamic physical alterations in membrane shape that are mediated in part by endophilin proteins. The endophilin B family of proteins contains an N-terminal Bin/amphiphysin/Rvs (N-BAR) domain that induces membrane curvature to regulate intracellular membrane dynamics. Whereas endophilin B1 (SH3GLB1/Bif-1) is known to be involved in a number of cellular processes, including apoptosis, autophagy, and endocytosis, the cellular function of endophilin B2 (SH3GLB2) is not well understood. In this study, we used genetic approaches that revealed that endophilin B2 is not required for embryonic development in vivo but that endophilin B2 deficiency impairs endosomal trafficking in vitro, as evidenced by suppressed endosome acidification, EGFR degradation, autophagic flux, and influenza A viral RNA nuclear entry and replication. Mechanistically, although the loss of endophilin B2 did not affect endocytic internalization and lysosomal function, endophilin B2 appeared to regulate the trafficking of endocytic vesicles and autophagosomes to late endosomes or lysosomes. Moreover, we also found that despite having an intracellular localization and tissue distribution similar to endophilin B1, endophilin B2 is dispensable for mitochondrial apoptosis. Taken together, our findings suggest that endophilin B2 positively regulates the endocytic pathway in response to growth factor signaling, autophagy induction, and viral entry.

philin B1 deficiency in mice promotes obesity (27), Alzheimer's disease pathology (28), and tumor development (19,21), indicating the importance of endophilin B1 in the maintenance of tissue homeostasis. Endophilin B2 was discovered by a yeast two-hybrid screen using endophilin B1 as a bait and exhibits 58% homology in amino acids and 67% homology in nucleotide sequence to endophilin B1 in humans (15). Despite its high sequence and structural similarity to endophilin B1 (29), the cellular function of endophilin B2 remains unclear.
In this study, we utilized genetic approaches to demonstrate that, whereas endophilin B2 is dispensable for embryonic development and mitochondrial apoptosis, loss of endophilin B2 impairs endosome acidification, leading to suppressed EGFstimulated EGF receptor (EGFR) degradation, autophagic flux, and influenza A virus entry.

Endophilin B2 is dispensable for embryonic development
To determine the importance of endophilin B2 in vivo, endophilin B2-deficient mice were generated from an embryonic stem cell clone containing a gene trap cassette in the first intron of the Sh3glb2 gene (Fig. 1A). Genotyping of the resultant mice demonstrated successful insertion of the cassette, which disrupted endophilin B2 protein expression in isolated mouse embryonic fibroblasts (MEFs) (Fig. 1, B and C). Similar to endophilin B1 (18), loss of endophilin B2 does not affect embryonic development, because pups were born with normal ratios of Mendelian inheritance and were indistinguishable from wildtype littermates. Moreover, endophilin B2 demonstrated a similar tissue expression profile to endophilin B1, as indicated by co-expression of endophilins B1 and B2 in the majority of tis- Figure 1. Endophilin B1 and B2 share similar expression profiles in mice. A, schematic of gene trap cassette insertion into endophilin B2 (B2) gene intron, genotyping primer design (described under "Experimental Procedures"), and cassette scheme (LTR, long terminal repeat; OV, Omnibank vector). B, genomic DNA prepared from mouse tail was subjected to PCR analysis using the primer sets shown in A. C, total cell lysates were prepared from MEFs with the indicated genotypes and subjected to immunoblotting using the indicated antibodies. B1, endophilin B1; B2, endophilin B2. D, tissue homogenates were prepared from 10-week-old endophilin B2KO and WT mice and subjected to immunoblotting using the indicated antibodies. Sk. muscle, skeletal muscle; WAT, white adipose tissue; BAT, brown adipose tissue. E, schematic of predicted murine endophilin B2 isoforms with predicted splicing and amino acid length. F, the mRNA expression profile for each endophilin B2 isoform was determined by semi-quantitative RT-PCR as described in supplemental Fig. S3. Relative expression of each isoform in each tissue was normalized to isoform 1 expression.
sues examined (Fig. 1D). Interestingly, because endophilin B2 assembles heterodimers with endophilin B1 (15), we found that loss of endophilin B2 suppressed endophilin B1 expression in several tissues, including the heart, stomach, spleen, lymph node, bladder, skeletal muscle, and brown adipose tissue, indicating that an epistatic interaction between endophilin B1 and B2 genes may regulate the stability of endophilin B1 in these tissues.
Furthermore, several tissue-specific isoforms of endophilin B1 have been implicated in alternative cellular functions in the brain and testis (30,31). Because at least three distinct bands for endophilin B2 were detected in brain tissue lysates ( Fig. 1D and supplemental Fig. S1), we sought to identify and characterize the isoforms of endophilin B2 in mice. The mouse endophilin B2 gene is located on chromosome 2 and contains 13 exons (32) that are predicted by the Ensembl database to encode five isoforms varying from 374 to 404 amino acids in length ( Fig. 1E and supplemental Fig. S2). To assess the tissue expression profile of the endophilin B2 isoforms, we performed semi-quantitative RT-PCR (supplemental Fig. S3 and Table S2). We observed that isoform 1 of endophilin B2 was the most ubiquitously expressed, whereas isoforms 2 and 4 were increased in the brain, and isoforms 3 and 5 were undetectable in the tissues that we examined ( Fig. 1F and supplemental Fig. S3).
Collectively, these results demonstrate that endophilin B2 is dispensable for murine embryonic development, shares a similar tissue distribution profile with endophilin B1, and exists as three predominant isoforms with isoform 1 ubiquitously expressed and isoforms 2 and 4 enriched in the brain.

Endophilin B2 is dispensable for mitochondrial apoptosis
Endophilin B1 was discovered as a pro-apoptotic BAX-interacting protein that promotes mitochondrial apoptosis (15,16,18). Because endophilin B1 and endophilin B2 display a similar tissue expression profile (Fig. 1D), are highly similar in protein structure, and heterodimerize (15,33), we first aimed to examine whether endophilin B2 also plays an important role in mitochondrial apoptosis. To this end, WT MEFs or MEFs lacking endophilin B1 (B1KO) or endophilin B2 (B2KO) were transduced with lentivirus encoding the BH3-only protein BCL-2modifying factor (BMF) at an equal multiplicity of infection (MOI). BMF sensitizes cells to apoptosis by directly antagonizing anti-apoptotic BCL-2 family proteins (BCL-2 (B-cell lymphoma 2), BCL-X L (B-cell lymphoma extra large), and MCL-1 (myeloid cell leukemia 1)) to allow for BAX and BAK (BCL-2 antagonist/killer) activation (34). Unexpectedly, unlike endophilin B1 deficiency, the loss of endophilin B2 did not enhance cell viability in MEFs transduced with BMF lentivirus ( Fig. 2A;   Figure 2. Endophilin B2 is dispensable for mitochondrial apoptosis. A, WT, endophilin B1 knock-out (B1KO), and endophilin B2 knock-out (B2KO) MEFs were transduced with lentiviruses encoding BMF at an MOI of 3 for 48 h and subjected to a sulforhodamine B (SRB) cell viability assay. B, WT and B2KO MEFs were treated with 5 g/ml ActD or 1 M STS for 48 h and subjected to an MTS cell viability assay. Data from two independent experiments performed in triplicate are shown in A and B. C, WT, B2KO, and B1KO HeLa cells were subjected to immunoblotting using the indicated antibodies. D, HeLa cells were treated with the indicated doses of ActD for 24 h and subjected to a PrestoBlue cell viability assay. Data represent three independent experiments performed in triplicate. E, HeLa cells were transduced with lentiviruses encoding BMF at an MOI of 3 for 24 h, stained with annexin V, and analyzed by flow cytometry. Data represent three independent experiments performed in duplicate. Statistical significance was determined by unpaired two-way ANOVA with Tukey's multiple comparison test (A, B, and D) and unpaired Student's t test with Holm-Sidak multiple-comparison correction (E). All values are mean Ϯ S.D. ns, not significant; **, p Յ 0.01; ***, p Յ 0.001.

Endophilin B2 plays a role in endosome maturation
To further explore the cellular function of endophilin B2, we examined the subcellular localization of endophilin B2. Immunofluorescence microscopy revealed that endophilin B2 resides on cytoplasmic structures, where it partially colocalizes with endophilin B1 (supplemental Fig. S4A). Subcellular fractionation analysis further revealed an enrichment of endophilin B2 and endophilin B1 in early (Rab5) and late (Rab7) endosomal fractions as well as in the microtubule-associated protein 1A/1B light chain 3 (LC3)-II-enriched autophagosomal fractions (supplemental Fig. S4B). To confirm endophilin B2 localization to endosomes, HeLa cells expressing C-terminal tagged monomeric DsRed endophilin B2 were treated with fluorescence-tagged EGF to stimulate endosomal trafficking and analyzed by immunofluorescence microscopy. Indeed, we observed that endophilin B2 localizes to EGF-positive, Rab5positive early endosomes and lysobisphosphatidic acid (LBPA)positive late endosome structures in response to EGF stimulation ( Fig. 3 and supplemental Fig. S6). Because endophilin B1 promotes endosomal acidification and trafficking (24,25,35), we hypothesized that endophilin B2 may have a similar role in endocytic processes.
To determine whether endophilin B2 regulates fluid-phase endocytosis, cells were pulsed with FITC-labeled dextran, and fluid-phase endocytosis was quantified by measuring the relative fluorescence intensity (36). In contrast to 5-(N-ethyl-Nisopropyl)-amiloride (EIPA), a selective inhibitor of sodiumhydrogen exchange that impairs fluid-phase endocytosis (37, 38) (supplemental Fig. S5), we did not observe any difference in the uptake of FITC-dextran between WT and endophilin B1-or B2-deficient HeLa cells (Fig. 4A). Thus, in contrast to endophilin A proteins, the endophilin B family of proteins do not appear to be required for plasma membrane internalization during fluid-phase endocytosis. Next, to assess whether endophilin B2 regulates the endocytic trafficking of internalized vesicles, we utilized dextran conjugated to the pH-sensitive pHrodo-Red dye. As the fluorescence intensity of this dye increases in response to decreasing pH, pHrodo-Red dextran monitors the acidification and maturation of endosomes (5). As expected, a time-dependent increase in the fluorescence intensity of pHrodo-Red dextran was observed upon pulse-chase of WT HeLa cells (Fig. 4B), indicating the progressive acidification of endosomes that reaches a plateau ϳ5 min after the addition of the dye. Interestingly, whereas pulse-chase of endophilin B1and B2-deficient cells with pHrodo-Red dextran resulted in a time-dependent increase in fluorescence intensity, the maximal fluorescence intensity was significantly reduced compared with WT cells ( Fig. 4C; WT to B2KO, p ϭ 0.0134; WT to B1KO, p ϭ 0.0459). Similarly, endophilin B2 depletion in MEFs did not affect the uptake of FITC-dextran ( Fig. 4D) but significantly reduced the maximal fluorescence intensity of pHrodo-Red dextran (Fig. 4, E and F; p ϭ 0.0322 (F)). Taken together, these data suggest that endophilin B2 functions in a similar manner to endophilin B1 to promote the acidification of endosomes.
Because endosomal acidification is critical for receptor degradation upon receptor-mediated endocytosis, we sought to determine the effect of endophilin B2 on the lysosomal degradation of EGFR. To this end, cells were serum-starved overnight to accumulate EGFR at the plasma membrane. Upon stimulation with EGF, EGFR is internalized (supplemental Fig.  S6) and trafficked to the lysosome for degradation with minimal receptor recycling to the plasma membrane (36,39). Using fluorescently tagged-EGF to stimulate and track EGFR (40), we observed that the loss of endophilin B2 significantly impairs the delivery of internalized EGFR to lysosomal-associated membrane protein 1 (Lamp1)-positive lysosomal structures (Fig. 5, A and B; p Ͻ 0.0001). Importantly, and in accordance with our results for fluid-phase endocytosis, the plasma membrane internalization of EGFR is independent of endophilin B2 (Fig. 5A).
Acidification is a critical part of endosomal maturation that allows for the activation of acid hydrolases along the endosomal-lysosomal pathway. Our results reveal that the loss of either endophilin B1 or B2 significantly delayed the degradation of EGFR upon EGF stimulation (Fig. 6, A and B; WT to B2KO, p Ͻ 0.0001; WT to B1KO, p ϭ 0.0002). Lysosomes are dynamic organelles that are re-formed after fusion events and require an acidic lumen for the acid hydrolases to properly degrade material (41). To determine whether endophilin B2 is important for the biogenesis and/or acidification of lysosomes, HeLa cells were stained with the cell-permeable lysosomal dye Lyso-Tracker and the pH-responsive lysosomal dye LysoSensor, respectively. Whereas LysoTracker signals were not significantly altered by the loss of endophilin B2, LysoSensor signals were surprisingly slightly enhanced in endophilin B2-deficient cells to indicate that the compartments are more acidic compared with WT ( Fig. 6C; p ϭ 0.0001). Collectively, these results suggest that suppressed EGFR degradation observed in endophilin B2 deficient cells is not indirectly caused by lysosome impairment.
Importantly, re-expression of endophilin B2-GFP but not GFP control was able to rescue EGFR degradation in endophilin B2-deficient cells (Fig. 7, A-C). To identify the domain(s) of endophilin B2 that regulate EGFR degradation, cells were also restored with expression of endophilin B2 mutants lacking the SH3 domain or the N-BAR domain (Fig. 7, A-C). Interestingly, we found that the N-BAR domain of endophilin B2 is required to rescue EGFR degradation in endophilin B2-deficient cells, whereas the SH3 domain appears to be dispensable (Fig. 7, B and C; B2KO ϩ GFP to B2KO ϩ B2-GFP, p ϭ 0.0030; B2KO ϩ GFP to B2KO ϩ B2-⌬SH3-GFP, p ϭ 0.0345; B2KO ϩ GFP to B2KO ϩ B2-⌬N-BAR-GFP, p ϭ 0.9897). Using co-immunoprecipitation, we confirmed that endophilin B2, indeed, het- C and F, maximum signal intensities in B and E were determined by curve fit analysis. Independent assays were normalized to WT signal, and then WT signals from the three independent assays were normalized to mean WT signal. Statistical significance was determined by unpaired one-way ANOVA with Tukey's multiple-comparison test (A and C) or unpaired Student's t test with Holm-Sidak multiple comparison correction (D and F). All values are mean Ϯ S.D. (error bars). ns, not significant; *, p Ͻ 0.05.

Endophilin B2 promotes endosome maturation and autophagic flux
erodimerizes with endophilin B1 (15) and that heterodimization requires the N-BAR domain (Fig. 7D). These data suggest that the membrane curvature-inducing activity of endophilin B2 and/or heterodimerization of endophilin B2 and endophilin B1 are important for the endocytic degradation of EGFR.
Taken together, these results indicate that endophilin B2 facilitates endosome maturation in both fluid-phase and receptor-mediated endocytic pathways. Further investigation is under way to dissect the precise role of endophilin B family proteins in endocytic trafficking.

Endophilin B2 deficiency decreases autophagic flux upon nutrient starvation
During autophagy, autophagosomes mature into degradative autophagic vacuoles by fusing with late endosomes and/or lysosomes (reviewed in Refs. 7 and 8). To determine whether the impairment in endosome maturation observed in endophilin B2-deficient cells affects autophagy, we performed an autophagic flux assay. During autophagic stimulation, such as nutrient starvation, cytosolic LC3-I is lipidated to form LC3-II bound to the autophagosome membrane (11). Autophagosome maturation occurs upon fusion with lysosomes and degradation of luminal contents; thus, measurement of the lysosomal turnover of LC3-II has been used as a reliable method to monitor autophagic flux (42,43). Because increases in LC3-II can indicate either an increase in autophagosome biogenesis or a block in lysosomal degradation, autophagic flux assays were performed in the presence or absence of the lysosomal inhibitor, bafilomycin A1 (Baf A1). We found that the starvationinduced lysosomal turnover of LC3-II was significantly sup- To further demonstrate the importance of endophilin B2 in autophagic flux, we performed the tandem-fluorescent LC3 (tf-LC3; mRFP-GFP-LC3) assay. The tf-LC3 assay discriminates between autophagosomes from autolysosomes (autophagosomes fused with lysosomes) due to the differences in pK a val-ues between RFP and GFP. Upon autophagosome and lysosome fusion, the GFP moiety of tf-LC3 becomes quenched by the acidic environment to result in RFP ϩ GFP Ϫ puncta (44). We found that nutrient starvation of HeLa cells increased RFP ϩ GFP ϩ puncta (autophagosomes) regardless of endophilin B2 expression, indicating that endophilin B2 is not required for autophagosome formation (Fig. 8E). In contrast, the generation of RFP ϩ GFP Ϫ puncta (autolysosomes) was reduced in nutrient-starved endophilin B2-deficient cells compared with WT (Fig. 8E), suggesting impaired delivery of autophagosomes to acidified late endosomes and/or lysosomes. Indeed, RFP ϩ GFP ϩ signals significantly accumulated in endophilin B2-deficient cells during nutrient starvation compared with WT ( Fig.  8F; p Ͻ 0.0001). Because Baf A1 slightly but significantly enhanced the abundance of RFP ϩ GFP ϩ puncta in nutrientstarved endophilin B2-deficient cells ( Fig. 8F; † † † †, p Ͻ 0.0001), a small degree of autophagic flux is maintained in the absence of endophilin B2. Nonetheless, these results clearly demonstrate an important role of endophilin B2 in autophagosome maturation.

Loss of endophilin B2 delays nuclear translocation and replication of influenza A viruses
To address the biological significance of the above results, we monitored influenza A virus infection in WT and endophilin B2-deficient cells. Influenza A virus is a member of the Orthomyxoviridiae family of enveloped viruses. Influenza A virus is internalized through receptor-mediated endocytosis and trafficked to late endosomes. The acidic environment (pH ϳ5.0) of late endosomes triggers conformational changes in the viral envelope hemagglutinin that result in the fusion of the endosomal and viral membranes, release and transport of viral ribonucleoprotein particle to the nucleus, and replication of the viral RNA genome (vRNA) in the nucleus (45)(46)(47)(48)(49)(50). Notably, influenza A is a pathogen that requires both endosomal acidification and deployment of autophagy to efficiently replicate in host cells (51). Because endophilin B2-deficient cells were observed to have a defect in endosome acidification, we hypothesized that endophilin B2 may regulate viral trafficking. To this end, we monitored the subcellular localization of endophilin B2 in MEFs during influenza A virus infection. Indeed, endophilin B2 is observed to colocalize with the influenza A vRNA-binding nucleoprotein (NP) in the cytoplasm at ϳ1-1.5 h post-infection (Fig. 9A), suggesting a role for endophilin B2 in viral trafficking. We next determined the effect of endophilin B2 depletion on vRNA trafficking and replication. In WT cells, we observed a time-dependent increase of NP in the nucleus and then in the cytoplasm, indicating replication of the viral genome and generation of new virions (Fig. 9B). Similar to fluidphase and receptor-mediated endocytosis, endophilin B2 was dispensable for the internalization of influenza A virus into host cells (Fig. 9B). However, both the nuclear and cytoplasmic accumulation of NP were significantly delayed by loss of endophilin B2 (Fig. 9C; p Ͻ 0.0001), suggesting that endophilin B2 is important for the efficient trafficking of vRNAs to the nucleus. Taken together, these data suggest that endophilin B2 deficiency effectively impairs the pH-dependent escape of influenza A virus from endosomes to suppress vRNA translocation

Endophilin B2 promotes endosome maturation and autophagic flux
into the nucleus for replication, a result consistent with impaired endosomal acidification upon the loss of endophilin B2.

Discussion
In this study, we demonstrate a novel role for endophilin B2 in endocytic membrane trafficking upon fluid-phase endocytosis, receptor-mediated endocytosis, starvation-induced autophagy, and influenza A viral infection. Notably, the loss of endophilin B2 appears to impair endosomal maturation, as supported by suppressed endosomal acidification during fluid-phase endocytosis, decreased lysosomal degradation of internalized EGFR, impaired starvation-induced autophagic flux, and delayed nuclear translocation of influenza A vRNA. Furthermore, our studies reveal that loss of endophilin B2 fails to efficiently deliver EGFR to Lamp1-positive compartments upon receptor-mediated endocytosis or newly generated autophagosomes to acidified late endosomes and/ or lysosomes. Because the loss of endophilin B2 does not impair internalization from the plasma membrane or lysosomal function, endophilin B2 appears to regulate endocytic membrane trafficking along the endosomal-lysosomal pathway.
In contrast, endophilin B2 has recently been reported to be indispensable for the internalization of ligand-stimulated glutamate receptors in cultured neuronal cells (52). We propose that the discrepancy in results may be due to the differential expression of endophilin B2 isoforms in neuronal and non-neuronal cells. Indeed, we have identified and characterized the murine endophilin B2 isoforms and reveal a tissue-specific increase of isoforms 2 and 4 in neuronal tissues compared with non-neuronal tissues that predominantly express isoform 1. Interestingly, whereas endophilin B1 promotes apoptosis in non-neuronal cells, brain-specific isoforms of endophilin B1 have been reported to have anti-apoptotic functions during ischemic injury (30).
Although we observed that endophilin B1 and endophilin B2 can assemble heterodimers and have a similar subcellular localization in endothelial and fibroblast cells, our studies provide insight into unique functions of each family member. Most notably, we revealed that endophilin B2 is dispensable for mitochondrial apoptosis. Although this result is not entirely surprising, given that endophilin B2 is not a direct interactor of BAX (15), our results suggest that endophilin B1 homodimers may be critical for its pro-apoptotic function. In contrast, the novel roles of endophilin B2 in endocytic trafficking, endosomal maturation, and autophagy are similar to previously defined functions of endophilin B1 in EGFR trafficking and degradation (24) and autophagic flux (19,21), suggesting that heterodimers of endophilin B proteins may regulate endocytic membrane dynamics. In support of this hypothesis, we observed the cofractionation of endophilin B2 with endophilin B1 in endosome-and autophagosome-enriched fractions and demonstrated that the N-BAR domain, but not the SH3 domain, is necessary for the heterodimerization of endophilin B1 and B2 as well as EGFR degradation. Moreover, loss of endophilin B2 leads to impairment of starvation-induced autophagic flux. Consistently, a similar effect on autophagic flux is observed in endophilin B1-deficient models (19,21), and a recent report has shown that endophilin B2 promotes the selective degradation of mitochondria by autophagy (mitophagy) through dimerization with endophilin B1 (33). Moreover, this finding is in agreement with the importance of endosome maturation for autophagic flux (7). Further studies are warranted to elucidate the interaction between endophilin B family homodimers and heterodimers in the regulation of intracellular membrane dynamics.
The precise molecular mechanism behind the regulation of endosome maturation by endophilin B2 is yet to be determined. As discussed above, endophilin B2 heterodimerization with The lines, boxes, and error bars represent median values, 25th to 75th percentiles, and 5th to 95th percentiles, respectively. Statistical significance was determined by two-way ANOVA with Tukey's multiple-comparison test. *, p Յ 0.05; ***, p Յ 0.001; **** and † † † †, p Ͻ 0.0001; ns, not significant. endophilin B1 may promote endosome maturation through the activation of phosphatidylinositol 3-kinase Vps34 and/ or the recruitment of the UVRAG (UV radiation resistanceassociated gene)-Hops complex (homotypic fusion and protein sorting complex) to early endosomes (19,53). In fact, our preliminary data show that UVRAG co-immunoprecipitates with endophilin B1 and endophilin B2 (data not shown). In addition, our data indicated that the N-BAR domain of endophilin B2 is necessary for endocytic degradative trafficking, whereas the SH3 domain is dispensable. This suggests that the membrane curvature functions and dimerization of endophilin B2 are necessary for endocytic trafficking. Alternatively, or in addition, we cannot eliminate the possibility that endophilin B2 may regulate endosomal maturation through its effects on the cytoskeleton. Recently, endophilin B2 has been identified as a regulator of the vimentin cytoskeletal network via its SH3-mediated interaction with the cytoskeletal linker protein plectin 1 (54).
Because the vimentin network controls late endosome and lysosome positioning as well as the acidification of endosomes and lysosomes (55), we propose that this interaction may be critical for the phenotype reported here. Moreover, vimentin has also been identified as a vital regulator for the trafficking and endosomal escape of influenza A viral ribonucleoprotein particle (56). Because the phenotype of vimentin-null cells infected with influenza A virus is similar to that observed in endophilin B2-deficient cells, we hypothesize that endophilin B2 may cooperate with vimentin to regulate endosome acidification and/or the positioning of endosomes during maturation.
Because endophilin B2-deficient mice are born at the expected Mendelian ratios and are indistinguishable from wildtype littermates, evidence of endophilin B2 in the pathogenesis of disease is currently lacking. However, high expression of endophilin B2 has been reported in aggressive human prostate tumors as well as primary tumors from a transgenic mouse The lines, boxes, and error bars represent median values, 25th to 75th percentiles, and 5th to 95th percentiles, respectively. Statistical significance was determined by two-way ANOVA with Tukey's multiple-comparison test. ns, not significant; ****, p Ͻ 0.0001. model of prostate cancer to suggest that endophilin B2 could serve as a potential biomarker (57)(58)(59). In direct opposition, decreased endophilin B1 expression is associated with cancer development and progression (60 -63). The opposite changes in expression observed for the endophilin B proteins in cancer have generated much speculation regarding the function of endophilin B2. On one hand, endophilin B2 has been proposed to have pro-apoptotic functions like endophilin B1 but to persist in an up-regulated state for unknown reasons. Alternatively, endophilin B2 has been suggested to be anti-apoptotic, such that its expression is positively selected for during tumor development. Because loss of endophilin B2 did not further sensitize cells to mitochondrial apoptosis in our studies, endophilin B2 may not be a direct suppressor of apoptosis (1). Interestingly, the identification of endophilin B2 as a novel promoter of endocytic trafficking may provide insight into its role in tumorigenesis, because dysregulated growth factor signaling and autophagy are hallmarks of cancer (64). Additionally, our findings implicate endophilin B2 in influenza A viral infection to provide a novel target for future investigation of viruses requiring an acidic pH for infectivity. Collectively, our studies provide a foundation to understand the physiological and pathological roles of endophilin B2 in human health and disease.

Generation of endophilin B2 knock-out mice
Endophilin B2-deficient mice with a 129/SvEv and C57BL/6 mixed genetic background were generated by the Texas A&M Institute for Genomic Medicine (Houston, TX) by the gene trap strategy using the OMNIBANK ES cell clone OST224737. This gene trap cassette includes a splice adaptor at the 5Ј end with neomycin selection and poly(A) termination signals followed by another ES cell-only promotor that expresses the first exon of the Bruton's tyrosine kinase gene with multiple termination codons to prevent expression of downstream fusion transcripts. The resultant mice were back-crossed with C57BL/6 mice (Jackson Laboratories, Bar Harbor, ME) more than 10 times to obtain endophilin B2KO mice with the genetic background used in this study. Genotyping was performed using standard PCR methods with the following primers: genomic forward Endophilin B2 promotes endosome maturation and autophagic flux primer (B2-F), 5Ј-GTTGGTGCTAATGGTTGCATCC-3Ј; genomic reverse primer (B2-R), 5Ј-AGATCCTAGCCTTCT-GACATCC-3Ј; and LTR-reverse (LTR-R), 5Ј-ATAAAC-CCTCTTGCAGTTGCATC-3Ј. All mice were maintained in accordance with federal guidelines, and studies were approved by the Pennsylvania State University Animal Care and Use Committee.

Semi-quantitative RT-PCR
Total RNA was isolated from tissues using TRIzol reagent (Thermo Fisher Scientific, catalog no. 15596026) according to the manufacturer's protocol. Quality of RNA was tested using the BioAnalyzer RNA 6000 nanochip (Agilent Technologies). cDNA was created from RNA using the High Capacity cDNA Reverse Transcriptase Kit (Thermo Fisher Scientific, catalog no. 4368814) according to the manufacturer's protocol. PCR amplification of endophilin B2 isoforms was achieved by using 2.5 ng of each cDNA using the primer pairs found in supplemental Table S1. Semi-quantitative PCR was performed using the GoTaq Colorless Master Mix (Promega, catalog no. M713) according to the manufacturer's protocol, using the scheme and protocol outlined in supplemental Fig. S2A and Table S2, respectively. Isoforms of endophilin B2 were quantified by the BioAnalyzer High Sensitivity chip (Agilent Technologies), and relative mRNA values of each isoform were calculated based on the ratios and differences of the PCR1 and -2 products.

Cell culture, viral transduction, and generation of stable cell lines
HeLa (CCL-2) and 293T/17 (HEK293T/17) cells were obtained from ATCC. SV40 large T antigen-immortalized MEFs were generated as described previously (21). All cells were maintained in DMEM supplemented with 10% FBS and 1% antibiotic-antimycotic solution (Corning, catalog no. 30-004-CI). Recombinant lentiviruses were produced and transduced to targeted cells as described previously (66). Lentiviral MOI was calculated from infectious viral units/ml and determined using the Lenti-X quantitative RT-PCR kit (Clontech, catalog no. 631235). To generate endophilin B2 knock-out HeLa cells, cells transduced with lentiviruses encoding Cas9 and endophilin B2 single-guide RNA were selected with 1.5 g/ml puromycin for 5 days followed by single-clone isolation by serial dilution. The resultant single clones (three individual clones) were pooled together and used for the experiments.

Fluorescence microscopy
Cells were fixed in 4% paraformaldehyde-PBS for 10 min, permeabilized with 100 g/ml digitonin for 10 min, and stained for the indicated antibodies unless otherwise noted. Fluorescence images were obtained using an Olympus IX81 deconvolution microscope (ϫ60 PLAN S-APO oil immersion objective, numeric aperture ϭ 1.35; ϫ40 PLAN FLUORITE oil immersion objective, numeric aperture ϭ 1.30) equipped with a Hamamatsu ORCA-R2 Digital CCD camera (catalog no. C10600-10B) or a Leica AOBS SP8 laser-scanning confocal microscope (ϫ40 PLAN-APOCHROMAT, numeric aperture ϭ 1.3 oil immersion objective), deconvolved using Slide-Book software (Intelligent Imaging Innovations) or Huygens deconvolution software (Scientific Volume Imaging), and analyzed using SlideBook software and/or Imaris software (Bitplane).

Endocytosis assays
To measure fluid-phase endocytosis and endosome acidification, cells were incubated in Live Cell Imaging Solution (Thermo Fisher Scientific, catalog no. A14291DJ) containing 50 g/ml FITC-dextran (10,000 Da; Sigma, catalog no. FD10S-100MG) or pHrodo Red-dextran (10,000 Da; Thermo-Fisher Scientific, catalog no. P10361), respectively, at 37°C for 5 min, rinsed twice with Live Cell Imaging Solution, and analyzed by spectrofluorometry using a CLARIOstar plate reader with MARS data management software (BMG Labtech). Fluorescence signals were normalized according to cell viability determined by PrestoBlue staining as described above. For Alexa Fluor 488-dextran (Thermo-Fisher Scientific, catalog no. D22910) treated with EIPA (Enzo Life Sciences, catalog no. ALX-550-266-M005), cells were either preincubated with 80 mM EIPA for 30 min or not and then incubated with 50 g/ml Alexa Fluor 488-dextran for 30 min with or without EIPA before fixation. Cell images were captured using the Olympus IX-81 microscope and processed using Slidebook version 6.0 software. EGF receptor-mediated endocytosis and endocytic trafficking were determined as described previously (24). Briefly, cells were serum-starved for 16 h; incubated in DMEM containing 20 mM Hepes (pH 7.5), 0.1% BSA (EMD Millipore, catalog no. 126575), and 100 ng/ml EGF (Thermo Fisher Scientific, catalog no. PHG0311) at 37°C for the indicated periods of time; and subjected to immunoblotting to measure EGFR degradation. To monitor receptor-mediated endocytosis, serumstarved cells were rinsed once with ice-cold PBS and then incubated in uptake medium (DMEM, 2% BSA, 20 mM Hepes, pH 7.5) containing 1 g/ml Alexa Fluor 488-EGF (Thermo Fisher Scientific, catalog no. E13345) on ice for 1 h. After washing three times with ice-cold PBS to remove unbound ligands, cells were incubated at 37°C for the indicated periods of time, fixed in 4% PFA-PBS for 10 min, and subjected to immunofluorescence microscopy. Colocalization reported as Manders' overlap coefficient was determined using the colocalization module in the Imaris version 8.1 software by manual thresholding.

Immunoprecipitation assay
HEK293T/17 cells were co-transfected by calcium phosphate co-precipitation with either empty GFP or one of the endophilin B2 mutant plasmids together with an endophilin B1-Myc tag vector. Upon overnight incubation post-transfection, cells were harvested and lysed in 1% Triton X-100 with 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, and 20 mM Tris-HCl (pH 7.5) containing protease and phosphatase inhibitors. Cell lysates (400 g of total protein) were immunoprecipitated using Myc-Trap_A beads (ChromoTek, catalog no. yta-10) and processed according to the manufacturer's protocol.

Autophagic flux assays
To determine lysosomal turnover of LC3-II, cells were incubated in serum-and amino acid-free Dulbecco's modified Eagle's medium (starvation medium; SM) (custom-ordered from Invitrogen) or control complete medium (CM) in the presence or absence of 100 nM Baf A1 (Enzo Life Sciences, catalog no. BML-CM110-0100) for 2 h and subjected to immunoblotting as described above. The levels of LC3-II were normalized to ␤-actin, and autophagic flux of LC3-II was calculated as follows: basal flux ϭ (CM with Baf A1) Ϫ CM; induced flux ϭ (SM with Baf A1) Ϫ SM, as defined previously (42).
For the mRFP-GFP-LC3 (tf-LC3) assay, cells were transduced with lentiviruses encoding tf-LC3 for 3 days, starved for 2 h, fixed in 4% PFA for 10 min, and subjected to fluorescence microscopy. Pearson's correlation coefficient was determined using SlideBook software.

Virus entry assay
The mouse-adapted influenza A virus H1N1 strain A/PR 8/34 (PR8) was produced in 10-day-old embryonated chicken eggs, purified by sucrose density gradient centrifugation, and quantitated by a fluorescence focus assay as described previously (67). To monitor viral entry, trafficking, and replication, MEFs seeded on coverslips at 150,000 cells/well of a 24-well plate were rinsed once with PBS and incubated in serum-free DMEM containing PR8 viruses for 1.5 h at 4°C. After rinsing twice with PBS, cells were further incubated in DMEM supplemented with 5% FBS at 37°C for the indicated periods of time, rinsed twice with PBS, fixed in 4% paraformaldehyde-PBS containing 0.1% Triton X-100 for 15 min, blocked in 1% normal goat serum for 60 min, stained for endophilin B2 and/or NP, and subjected to fluorescence deconvolution or confocal microscopy. Nuclei were stained with DAPI to generate nuclear masks, and NP fluorescence signals in the nucleus were quantified using SlideBook software.

Statistical analyses
Multiple independent experiments were tested for statistical significance using GraphPad Prism version 7.0. Statistical tests performed include Student's t test or ANOVA as outlined in the figure legends with appropriate post-hoc tests when necessary. The threshold for statistical significance for each test was set at 95% confidence (p Ͻ 0.05).