Genetic Evidence for a Mammalian Retromer Complex Containing Sorting Nexins 1 and 2
Total Page:16
File Type:pdf, Size:1020Kb
Genetic evidence for a mammalian retromer complex containing sorting nexins 1 and 2 Courtney T. Griffin*, JoAnn Trejo†, and Terry Magnuson*‡ *Department of Genetics and Carolina Center for Genome Sciences and †Department of Pharmacology, University of North Carolina, Chapel Hill, NC 27599 Edited by David D. Sabatini, New York University School of Medicine, New York, NY, and approved September 8, 2005 (received for review December 21, 2004) We have previously shown that the putative mammalian retromer deficient (Vps26tm1Cos) mice were a gift from Frank Costantini components sorting nexins 1 and 2 (Snx1 and Snx2) result in (Columbia University, New York) and were identical in pheno- embryonic lethality when simultaneously targeted for deletion in type to those generated by a previously described gene trap, mice, whereas others have shown that H58 (also known as which was presumed to disrupt H58 (10, 11). All mice were mVps26), another retromer component, results in similar lethality maintained on a mixed genetic background. For embryonic when targeted for deletion. In the current study, we address the analysis, noon on the day of plug detection was counted as genetic interaction of these mammalian retromer components in embryonic day 0.5 (E0.5). mice. Our findings reveal a functional interaction between H58, SNX1, and SNX2 and strongly suggest that SNX2 plays a more Genotyping. The Snx1 and Snx2 wild-type and targeted alleles critical role than SNX1 in retromer activity during embryonic were detected as described in ref. 9, except that PCRs for the development. This genetic evidence supports the existence of different alleles were run separately. Genotyping primers and mammalian retromer complexes containing SNX1 and SNX2 and PCR conditions for H58 amplification are described in Sup- identifies SNX2 as an important mediator of retromer biology. porting Methods, which is published as supporting information on Moreover, we find that mammalian retromer complexes contain- the PNAS web site. ing SNX1 and SNX2 have an essential role in embryonic develop- ment that is independent of cation-independent mannose 6-phos- Reverse Transcription (RT)-PCR. Total RNA was generated from a GENETICS phate receptor trafficking. litter of E8.5 wild-type CD-1 embryos or their yolk sacs by using TRIzol reagent (Invitrogen) and were reverse transcribed, as embryonic lethality ͉ H58 ͉ vacuolar protein sorting ͉ cation-independent described in Supporting Methods. One microliter of each RT mannose 6-phosphate receptor ͉ yolk sac reaction was amplified in a 50-l PCR by using primers described in Supporting Methods. RT-PCRs were performed under iden- membrane coat retromer complex mediates endosome-to- tical conditions for 25 cycles and with a 55°C annealing temper- AGolgi trafficking of the vacuolar hydrolase receptor Vps10p ature. PCR products were analyzed with MULTIANALYST soft- in yeast (1). Retromer complexes are comprised of Vps35p, ware (Bio-Rad). EST IMAGE clones nos. 578066 (Snx1) and Vps29p, Vps26p, Vps5p, and Vps17p. Mammalian orthologs for 891452 (Snx2) were used as cDNA templates to compare the each of these proteins, except Vps17p, have been identified, PCR amplification ability of the Snx1 and Snx2 primers used in thereby suggesting that mammalian cells may use a trafficking the RT-PCRs. complex similar in molecular composition to the yeast retromer complex. Furthermore, recent studies employing cell lines indi- Cell Lines and Assays. Primary mouse embryonic fibroblasts cate that mammalian retromer complexes participate in traffick- (MEFs) were isolated as described in ref. 9, except that H58Ϫ/Ϫ ing of the cation-independent mannose 6-phosphate receptor embryos were dissected at E8.5. We established permanent, (CI-MPR), the functional homolog of Vps10p, suggesting that monomorphic, contact-inhibited MEF cell lines by subculturing the function of retromer complexes may also be conserved cells continuously until they emerged from a crisis period of slow between yeast and mammals (2–4). growth (Ϸ25 passages) as described in ref. 12 and 13. Immuno- Sorting nexin (SNX) 1 and SNX2 are 63% identical at the fluorescence analysis was performed as described in ref. 9 by amino acid level and are both mammalian orthologs of Vps5p using polyclonal (anti-rat) anti-CI-MPR antibodies provided by (5). Despite their homology to Vps5p, the function of SNX1 and Nancy Dahms (Medical College of Wisconsin, Milwaukee, WI) SNX2 in mammalian cells is poorly understood. Originally and by J. Paul Luzio (University of Cambridge, Cambridge, discovered in a screen designed to identify molecules involved in U.K.), a monoclonal anti-EEA1 antibody (BD Biosciences), and lysosomal sorting of the epidermal growth-factor receptor (6), AlexaFluor-conjugated secondary antibodies (Molecular SNX1 has also been shown to associate with a number of Probes). Images were collected by using an Olympus DSU different receptors as well as other putative mammalian retro- spinning disk confocal microscope configured with an IX71 mer components, including SNX2, in cultured cell lines (5, 7, 8). fluorescent microscope fitted with a PlanApo ϫ60 oil objective However, the functional implications of these interactions re- and Hamamatsu ORCA-ER digital camera. Fluorescent images main obscure, and it has been difficult to differentiate between of X-Y sections at 0.15 m were collected sequentially by using potential roles for SNX1 and SNX2 in retromer complexes Intelligent Imaging Innovations SLIDEBOOK 4.1 software. The versus roles for these sorting nexins separable from retromer final composite images were created by using PHOTOSHOP activity. In the present study, we sought to define the participa- (Adobe Systems, San Jose, CA). Western blots for CI-MPR tion and activity of SNX1 and SNX2 in mammalian retromer detection were performed on lysates prepared from untreated complexes through genetic analyses in mice. We provide genetic cells or cells treated for 17 h with 40 g͞ml cycloheximide evidence for the involvement of these sorting nexins in mam- malian retromer complexes and reveal an essential role for SNX2 in retromer activity during embryonic development. This paper was submitted directly (Track II) to the PNAS office. Abbreviations: CI-MPR, cation-independent mannose 6-phosphate receptor; En, embry- Methods onic day n; MEF, mouse embryonic fibroblast; RT, reverse transcription; SNX, sorting nexin. Animals. Generation of Snx1-deficient (Snx1tm1Mag) and Snx2- ‡To whom correspondence should be addressed. E-mail: terry[email protected]. deficient (Snx2tm1Mag) mice was described in ref. 9. H58- © 2005 by The National Academy of Sciences of the USA www.pnas.org͞cgi͞doi͞10.1073͞pnas.0409558102 PNAS ͉ October 18, 2005 ͉ vol. 102 ͉ no. 42 ͉ 15173–15177 Downloaded by guest on September 23, 2021 Table 1. Snx2؊͞؊;H58؉͞؊ embryos result in 90% lethality Table 2. Snx1؉͞؊;Snx2؊͞؊;H58؉͞؊ embryos result in 100% lethality Live progeny Genotype (expected progeny) Live progeny Genotype (expected progeny) Snx2ϩ͞Ϫ;H58ϩ͞ϩ 77 (51.25) Snx2ϩ͞Ϫ;H58ϩ͞Ϫ 61 (51.25) Snx1ϩ͞Ϫ;Snx2ϩ͞ϩ;H58ϩ͞ϩ 29 (25) Snx2Ϫ͞Ϫ;H58ϩ͞ϩ 61 (51.25) Snx1ϩ͞Ϫ;Snx2ϩ͞ϩ;H58ϩ͞Ϫ 30 (25) Snx2Ϫ͞Ϫ;H58ϩ͞Ϫ 6 (51.25) Snx1ϩ͞Ϫ;Snx2ϩ͞Ϫ;H58ϩ͞ϩ 52 (50) Snx1ϩ͞Ϫ;H58ϩ͞ϩ 22 (20.25)* Snx1ϩ͞Ϫ;Snx2ϩ͞Ϫ;H58ϩ͞Ϫ 53 (50) Snx1ϩ͞Ϫ;H58ϩ͞Ϫ 15 (20.25)* Snx1ϩ͞Ϫ;Snx2Ϫ͞Ϫ;H58ϩ͞ϩ 16 (15) Snx1Ϫ͞Ϫ;H58ϩ͞ϩ 25 (20.25)* Snx1ϩ͞Ϫ;Snx2Ϫ͞Ϫ;H58ϩ͞Ϫ 0 (15) Snx1Ϫ͞Ϫ;H58ϩ͞Ϫ 19 (20.25)* Snx1ϩ͞Ϫ;Snx2ϩ͞Ϫ;H58ϩ͞ϩ 27 (23.5)* Snx1ϩ͞Ϫ;Snx2ϩ͞Ϫ;H58ϩ͞Ϫ 19 (23.5)* Shown are the number of genotypes of live progeny from 28 litters gen- Snx1Ϫ͞Ϫ;Snx2ϩ͞Ϫ;H58ϩ͞ϩ 28 (23.5)* erated by Snx2ϩ͞Ϫ;H58ϩ͞Ϫ ϫ Snx2Ϫ͞Ϫ or by Snx2Ϫ͞Ϫ;H58ϩ͞Ϫ ϫ Snx2ϩ͞Ϫ Ϫ͞Ϫ ϩ͞Ϫ  ϩ͞Ϫ matings and 11 litters generated by Snx1ϩ͞Ϫ;H58ϩ͞Ϫ ϫ Snx1Ϫ͞Ϫ or by Snx1Ϫ͞Ϫ; Snx1 ;Snx2 ;H 58 20 (23.5)*  ϩ͞Ϫ ϫ ϩ͞Ϫ Ϫ͞Ϫ  ϩ͞Ϫ H 58 Snx1 crosses. Note the 90% lethality for Snx2 ;H 58 Shown are the number of genotypes of live progeny from 24 litters Ͻ 2 animals (P 0.001 by test). *, the number of genotypes of live progeny from generated by Snx2ϩ͞Ϫ;H58ϩ͞Ϫ ϫ Snx1Ϫ͞Ϫ;Snx2ϩ͞Ϫ matings and 11 litters ϩ/Ϫ  ϩ/Ϫ ϫ Ϫ/Ϫ Ϫ/Ϫ  ϩ/Ϫ ϫ 11 litters generated by Snx1 ;H 58 Snx1 or by Snx1 ;H 58 generated by Snx1Ϫ͞Ϫ;H58ϩ͞Ϫ ϫ Snx1ϩ͞Ϫ;Snx2Ϫ͞Ϫ matings. Note the com- ϩ/Ϫ Ϫ͞Ϫ  ϩ͞Ϫ Snx1 matings. No significant lethality was detected for Snx1 ;H 58 plete lethality for Snx1ϩ͞Ϫ;Snx2Ϫ͞Ϫ;H58ϩ͞Ϫ animals (P Ͻ 0.005 by 2 test). The animals. expected numbers account for 40% lethality of Snx1ϩ͞Ϫ;Snx2Ϫ͞Ϫ embryos, as reported in ref. 9. *, the number of genotypes of live progeny from 11 litters generated by Snx1Ϫ/Ϫ;H58ϩ/Ϫ ϫ Snx1ϩ/Ϫ;Snx2Ϫ/Ϫ matings. No significant (Sigma). Western blots for mVPS35 detection were performed lethality was detected for Snx1Ϫ͞Ϫ;Snx2ϩ͞Ϫ;H58ϩ͞Ϫ animals. by using a polyclonal (anti-human) anti-Vps35 antibody (5) provided by Carol Haft (National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda). Membranes were Snx1ϩ/Ϫ matings (Table 1), indicating that SNX1 has a weaker probed with an anti-actin antibody (Santa Cruz Biotechnology) interaction with H58 and the retromer complex compared with to control for equal loading of lysates. Immunoblots were SNX2. developed with ECL-plus (Amersham Pharmacia), imaged by Although Snx1Ϫ/Ϫ;Snx2ϩ/Ϫ mice are viable, 40% of autoradiography, and quantified by using a Fluor-S Imager Snx1ϩ/Ϫ;Snx2Ϫ/Ϫ embryos were previously reported to die (Bio-Rad). during development, indicating that SNX1 and SNX2 are not completely functionally redundant (9). Strikingly, we Yolk Sac and Visceral Endoderm Immunostaining. For analysis of found that no Snx1ϩ/Ϫ;Snx2Ϫ/Ϫ;H58ϩ/Ϫ mice generated from CI-MPR localization in yolk sac cells, embryos and extraembry- Snx2ϩ/Ϫ;H58ϩ/Ϫ ϫ Snx1Ϫ/Ϫ;Snx2ϩ/Ϫ matings survived develop- onic yolk sacs were dissected at E8.5, and yolk sacs were ment (Table 2).