In contrast, the vacuoles within unc-32 and umps-1 embryos contained birefringent material Table 1; Fig.. Prominent vacuoles visible with DIC microscopy that were lacking in wild type A
Trang 1enlarged lysosome-related organelles in embryos lacking umps-1 function
Steven Levitte1, Rebecca Salesky1, Brian King2, Sage Coe Smith2, Micah Depper1, Madeline Cole2 and Greg J Hermann1,2
1 Department of Biology, Lewis & Clark College, Portland, OR, USA
2 Program in Biochemistry and Molecular Biology, Lewis & Clark College, Portland, OR, USA
Introduction
Lysosome-related organelles (LROs) represent a
diverse collection of specialized compartments that
share features in common with conventional lysosomes
[1–3] LROs perform a variety of important physiologi-cal functions In mammals, for example, lamellar bodies function in the storage and release of lung
Keywords
gut granule; lysosome-related organelle;
orotic aciduria; UMPS
Correspondence
G Hermann, Department of Biology, Lewis
& Clark College, 0615 S.W Palatine Hill Rd,
Portland, OR 97219, USA
Fax: +1 503 768 7658
Tel: +1 503 768 7568
E-mail: hermann@lclark.edu
(Received 16 October 2009, revised 26
December 2009, accepted 5 January
2010)
doi:10.1111/j.1742-4658.2010.07573.x
Gut granules are cell type-specific lysosome-related organelles found within the intestinal cells of Caenorhabditis elegans To investigate the regulation
of lysosome-related organelle size, we screened for C elegans mutants with substantially enlarged gut granules, identifying alleles of the vacuolar-type
H+-ATPase and uridine-5¢-monophosphate synthase (UMPS)-1 UMPS-1 catalyzes the conversion of orotic acid to UMP; this comprises the two ter-minal steps in de novo pyrimidine biosynthesis Mutations in the ortholo-gous human gene UMPS result in the rare genetic disease orotic aciduria The umps-1()) mutation promoted the enlargement of gut granules to 250 times their normal size, whereas other endolysosomal organelles were not similarly affected UMPS-1::green fluorescent protein was expressed in embryonic and adult intestinal cells, where it was cytoplasmically localized and not obviously associated with gut granules Whereas the umps-1()) mutant is viable, combination of umps-1()) with mutations disrupting gut granule biogenesis resulted in synthetic lethality The effects of mutations
in pyr-1, which encodes the enzyme catalyzing the first three steps of de novo pyrimidine biosynthesis, did not phenotypically resemble those of umps-1()); instead, the synthetic lethality and enlargement of gut granules exhibited by the umps-1()) mutant was suppressed by pyr-1()) In a search for factors that mediate the enlargement of gut granules in the umps-1()) mutant, we identified WHT-2, an ABCG transporter previously implicated in gut granule function Our data suggest that umps-1()) leads
to enlargement of gut granules through a build-up of orotic acid WHT-2 possibly facilitates the increase in gut granule size of the umps-1()) mutant
by transporting orotic acid into the gut granule and promoting osmotically induced swelling of the compartment
Abbreviations
DAPI, 4¢,6-diamidino-2-phenylindole; DIC, differential interference contrast; GFP, green fluorescent protein; HPS, Hermansky–Pudlak syndrome; LRO, lysosome-related organelle; ODC, orotidine-5¢-monophosphate decarboxylase; OMP, orotidine 5¢-monophosphate; OPRT, orotate phosphoribosyltransferase; RNAi, RNA interference; UMPS, uridine-5¢-monophosphate synthase; V-ATPase, vacuolar-type
H + -ATPase.
Trang 2surfactant, and melanosomes act to synthesize and
store body pigments [2,4] Investigation of the genetic
basis of Hermansky–Pudlak syndrome (HPS), which is
characterized by defects in the formation and function
of LROs, has led to the identification of 15 genes
act-ing in the traffickact-ing pathways to LROs [5]
Although mutations in HPS genes typically result in
a reduced number of LROs, there is a subset of HPS
mutations that additionally promote the formation of
dramatically enlarged LROs, including melanosomes
[6,7] and lamellar bodies [8,9] Lamellar bodies are
similarly enlarged in Tangier disease, which results
from defects in the function of the ABC transporter
ABCA1 [10] Most dramatically, nearly every class of
LRO is enlarged in patients with Chediak–Higashi
syn-drome [11] In none of these diseases do we clearly
understand the mechanistic basis for LRO
enlarge-ment, reflecting our lack of insight into the processes
that control the stereotypic size and morphology of
LROs
Gut granules are intestinal cell-specific LROs found
in the nematode Caenorhabditis elegans In addition to
typical lysosomal characteristics, gut granules stain
with Nile Red, a marker for hydrophobic material,
and contain birefringent and autofluorescent materials,
which are uniquely localized to the gut granule
[12–16] Gut granule formation is initiated during early
embryogenesis, soon after endoderm specification, and
intestinal cells typically contain hundreds of gut
granules [12–14] Gut granule biogenesis requires the
activity of conserved genes that function generally in
LRO formation, including those encoding the HOPS
complex, the AP-3 complex, the ABC transporter
PGP-2, and the Rab GTPase GLO-1 [13,15]
Here we describe the results of a genetic screen to
identify factors involved in regulating gut granule size,
and present a phenotypic, cellular and molecular
characterization of one of these genes, umps-1
Results
A screen for mutants with enlarged gut granules
in embryonic intestinal cells
Gut granules are abundant, cell type-specific, LROs
that are present within the intestinal cells of C elegans
embryos, larvae, and adults [12,13] The formation of
gut granules is initiated during early embryogenesis,
and is directly controlled by the regulatory program
governing intestinal cell fate and differentiation in the
early C elegans embryo [14,17,18] We have been
investigating the mechanisms controlling the assembly
and morphology of gut granules during embryogenesis
in order to identify the primary regulators of these processes
In adult C elegans intestinal cells, the enlargement
of endolysosomal organelles typically results in a Vac (vacuolated appearance) phenotype, characterized by the presence of cytoplasmic vacuoles when visualized with differential interference contrast (DIC) micros-copy The vacuolization of the adult intestine is associ-ated with enlargement of early endosomes [19], recycling endosomes [20], and late endosomes⁄ lyso-somes [21,22] We reasoned that enlargement of gut granules would similarly result in a Vac phenotype
We first analyzed strains known to exhibit enlarged endolysosomal compartments in adult or embryonic intestinal cells for vacuolization of the embryonic intestine Only one of the mutants, ppk-3()), displayed
an embryonic Vac phenotype (Fig 1D,E; Table 1) We therefore performed a screen for additional mutants that contained vacuoles within embryonic intestinal cells
We identified seven mutants exhibiting a Vac pheno-type Complementation tests and molecular cloning showed that these mutants were defective in three genes: ppk-3 (one allele), unc-32 (five alleles), and umps-1 (one allele) The ppk-3()) mutant displayed prominent vacuoles in the intestine (Fig 1D,E; Table 1), as well as in other embryonic cells, as has been reported previously [21] The ppk-3 gene encodes
a phosphatidylinositol-3-kinase that catalyzes the for-mation of phosphatidylinositol 3,5-bisphosphate and is orthologous to PIKfyve in mammals and Fab1p in yeast [21] Cells lacking the function of these kinases display dramatically enlarged late endolysosomal com-partments [23] The unc-32()) and umps-1()) mutants contained vacuoles exclusively within embryonic intes-tinal cells (Fig 1G,H,J,K; Table 1) The unc-32 gene encodes an intestinally expressed V0 subunit of the vacuolar-type H+-ATPase (V-ATPase) [24,25] The V-ATPase associates with embryonic gut granules [13], where it functions in acidification [26] The umps-1 gene encodes UMP synthase (UMPS), which is pre-dicted to function in de novo pyrimidine biosynthesis [27]
We analyzed whether gut granules were enlarged in the vac mutants Embryonic gut granules contain bire-fringent material [13,14] and the integral membrane ABC transporter PGP-2 [15] The vacuoles within ppk-3()) intestinal cells did not contain birefringent material (Fig 1D,E; Table 1) Although PGP-2-marked gut granules were slightly enlarged in ppk-3()) embryos (Fig 1F; Fig S1), they did not match the size of vacu-oles present within ppk-3()) embryos (Fig 1D,F) These observations indicate that the vacuoles visible in
Trang 3ppk-3()) embryos by DIC microscopy are not gut
granules The ppk-3()) adults displayed a slight
enlargement of autofluorescent gut granules (Fig S1)
Thus, ppk-3 plays only a minor role in regulating gut
granule size In contrast, the vacuoles within unc-32())
and umps-1()) embryos contained birefringent material
(Table 1; Fig 1G,H,J,K), and dramatically enlarged
PGP-2-containing compartments were present in
these mutants (Fig 1I,L), consistent with gut granule
enlargement Here, we present our analysis of the role
that UMPS-1 plays in gut granule formation and
morphology; detailed studies of the role that the
V-AT-Pase plays in regulating these processes will be described
elsewhere
Disrupting the activity of umps-1, a gene that
functions in pyrimidine biosynthesis, leads to a
Vac phenotype
We identified umps-1 as the gene disrupted in the vac
mutant zu456 (see Experimental procedures)
Promi-nent vacuoles were present within the intestinal cells of
umps-1(zu456) embryos from the ‘lima bean’ stage
through to hatching (Fig 1J; Fig S2) Vacuoles
dimin-ished in size and number during the L1 stage, and
L2-stage to adult-stage animals exhibited normal
intes-tinal morphology (Fig S2) umps-1(RNAi) led to an
embryonic Vac phenotype that was indistinguishable
from that caused by umps-1(zu456) (Table 1) Despite the dramatic vacuolization of the embryonic intestine, umps-1(zu456) animals can be maintained as a homo-zygous line
UMPS-1 is orthologous to mammalian UMPS [27],
a bifunctional enzyme that catalyzes the two terminal reactions in de novo pyrimidine biosynthesis [28] (Fig 2A,C) The orotate phosphoribosyltransferase (OPRT) activity of UMPS promotes the conversion of orotic acid to orotidine 5¢-monophosphate (OMP) The OMP decarboxylase (ODC) activity of UMPS catalyzes the formation of UMP from OMP The
C elegans UMPS-1 protein exhibits both OPRT and ODC activity in vitro [27] The sequence of umps-1 from zu456 showed a mutation that destroys the pre-dicted translation initiation site (Fig 2B) Use of the next downstream ATG would result in the formation
of a short, out-of-frame peptide We therefore con-clude that zu456 is probably a null allele of umps-1 The C elegans gene R12E2.11 codes for a protein homologous to the OPRT domain of human and
C elegansUMPS In vitro, R12E2.11 has OPRT activ-ity but lacks ODC activactiv-ity [27], suggesting that it might functionally overlap with UMPS-1 R12E2.11(RNAi) did not result in the formation
of embryonic vacuoles (Table 1) In addition, R12E2.11(RNAi) did not obviously alter the forma-tion and size of embryonic vacuoles in umps-1(zu456)
Fig 1 Analysis of embryonic vac mutants for gut granule enlargement Prominent vacuoles visible with DIC microscopy that were lacking in wild type (A) were present within the intestinal cells of ppk-3( )) (D), unc-32( )) (G) and umps-1()) (J) embryos The vacuoles (white arrowheads) (D) in ppk-3( )) embryos did not contain birefrin-gent material (white arrows) (E), as they did
in unc-32( )) and umps-1()) embryos (G, H,
J, K) (C, F, I, L) PGP-2 staining (marked by white arrows) in pretzel-stage embryos PGP-2-labeled compartments in ppk-3( )) embryos (F) were slightly enlarged in comparison with the wild type (C) In contrast, PGP-2-containing compartments were dramatically enlarged in unc-32( )) (I) and umps-1( )) (L) embryos The intestine is flanked by black arrowheads Embryos are approximately 50 lm in length.
Trang 4embryos (Table 1) or the persistence of intestinal
vacu-oles in umps-1(zu456) larvae (data not shown),
sug-gesting that R12E2.11 does not play a major role in
regulating the size of intestinal organelles Moreover,
R12E2.11(RNAi) did not result in phenotypes
charac-teristic of defects in pyrimidine biosynthesis, suggesting
that it is not essential for this process (Table S1 and data not shown)
The umps-1(zu456) line, while being viable, exhib-ited partially penetrant embryonic and larval lethality Fifty-six per cent of umps-1(zu456) embryos failed to hatch (Table S2) In addition, 30% of umps-1(zu456)
Table 1 Vacuole formation in embryonic intestinal cells All strains were grown at 22 C Pretzel-stage embryos were scored using DIC microscopy for the presence of vacuoles in embryonic intestinal cells Large vacuoles were typically ‡ 1.5 lm, and small vacuoles were between 0.8 and 1.4 lm in diameter Polarization microscopy was used to assess the presence of birefringent material within vacuoles.
n, number of embryos scored.
Genotype
Percentage of embryos with large vacuoles containing birefringent material
Percentage of embryos with large vacuoles lacking birefringent material
Percentage of embryos with small vacuoles containing birefringent material n
Enlarged endolysosomal compartments
V-ATPase
De novo pyrimidine biosynthesis
Transgenic rescue g
Double mutantsh
Mosaic RNAi
Trang 5L1-stage larvae did not reach adulthood (Table S2).
We found that the overall rate of embryogenesis was
delayed in umps-1(zu456) embryos; however, all of the
major tissues appeared to be properly specified and to
differentiate normally, and there were no obvious
developmental defects in umps-1(zu456) embryos prior
to the bean stage (data not shown) To determine when embryogenesis was affected in umps-1()) embryos, we monitored the development of individual bean-stage umps-1(zu456) and wild-type embryos Thirty-five per cent (n = 49) of umps-1(zu456) embryos elongated four-fold, whereas 100% of
Table 1 (Continued)
Genotype
Percentage of embryos with large vacuoles containing birefringent material
Percentage of embryos with large vacuoles lacking birefringent material
Percentage of embryos with small vacuoles containing birefringent material n
a Embryos scored were the progeny of cup-5(zu223) unc-36(e251) adults derived from a cup-5(zu223) unc-36(e251) ⁄ qC1 line b Embryos scored were the progeny of + ⁄ szT1[lon-2(e678)]; 3(ok1150) ⁄ szT1 adults Twenty-five per cent of the embryos were predicted to be ppk-3( )) ⁄ ppk-3()) c
The 32 alleles analyzed result in zygotic lethality Therefore, the embryos scored were the progeny of dpy-17(e164) unc-32( )) ncl-1(e1865) ⁄ qC1 dpy-19(e1259) glp-1(q339) adults Twenty-five per cent of the embryos were predicted to be unc-32()) ⁄ unc-32()) The linked dpy17(e164) ncl-1(e1865) markers did not result in a vacuole phenotype d The embryos scored were the progeny of umps-1(+) ⁄ umps-1( )) adults Twenty-five per cent of the embryos were expected to be umps-1()) ⁄ umps-1()) e
umps-1(+); mIs11[GFP] males were mated with umps-1(zu456) hermaphrodites, and outcross umps-1( )) ⁄ umps-1(+) embryos were recognized by their GFP expression and scored f The wild type or the indicated strain was grown on plates containing E coli expressing dsRNA against the listed gene g Embryos from parents containing extrachromosomal arrays were scored Owing to lack of segregation of the arrays, not all of the progeny will inherit the transgene [78], so some embryos from parents containing WRM0627dD02 still exhibit the umps-1( )) phenotype Only embryos expres-sing GFP, and therefore having inherited the UMPS-1::GFP array, were scored for intestinal vacuoles h Of the single mutants ⁄ RNAi exam-ined in the double mutant analysis, only umps-1( )) single mutants result in the formation of vacuoles within intestinal cells i
Embryos scored were the progeny of umps-1( )); apt-6()) parents, which exhibit 100% maternal effect lethality j
Embryos scored were the progeny
of umps-1( )) ⁄ umps-1()); glo-1()) ⁄ glo-1(+) parents The umps-1()); glo-1()) embryos were identified by the loss of the birefringent material phenotype exhibited by glo-1( )) embryos [13] k Embryos scored were the progeny of umps-1( )) ⁄ umps-1()); pgp-2()) ⁄ pgp-2(+) parents Twenty-five per cent of the embryos were expected to be umps-1( )); pgp-2()) The double mutants were identified by the loss or reduction
in the amount of birefringent material exhibited by pgp-2( )) homozygotes l Embryos scored were the progeny of umps-1( )) ⁄ umps-1()); wht-2( )) ⁄ wht-2(+) parents Twenty-five per cent of the embryos were expected to be umps-1()); wht-2()) m pyr-1(cu8) embryos exhibited reces-sive maternal effect suppression of umps-1(zu456). nThe strain was scored when grown on plates expressing F33E2.4-derived dsRNA F33E2.4 is not required for proper gut granule formation or morphology.
W
A
B
C
Fig 2 zu456 disrupts the activity of the bifunctional enzyme UMPS-1, which func-tions in de novo pyrimidine biosynthesis (A) The C elegans UMPS-1 protein contains distinct domains that mediate its OPRT and ODC activities (B) zu456 alters the pre-dicted translation initiation site of umps-1 (underlined in bold); use of the next poten-tial downstream start codon results in the formation of a short out-of-frame peptide (C) The pathway of de novo pyrimidine bio-synthesis in C elegans The proteins that catalyze each reaction are listed beside the arrows.
Trang 6wild-type embryos (n = 19) did so Thirty-five per cent
(n = 49) of umps-1(zu456) embryos arrested at
vari-ous stages between the bean stage and four-fold stage
of elongation: 10% arrested at the bean stage, 8%
arrested between the 1.5-fold stage and two-fold stage,
12% arrested between the two-fold stage and
three-fold stage, and 5% arrested between the three-three-fold
stage and four-fold stage Interestingly, we found that
30% (n = 49) of umps-1(zu456) embryos lysed,
typi-cally prior to elongation Lysis probably results from
umps-1()) embryos being sensitive to the mechanical
pressure associated with placing embryos between a
3% agarose pad and a coverslip These observations
indicate that umps-1()) activity is important for
embryonic and larval development, and the arrest and
lysis phenotypes suggest that umps-1(zu456)
compro-mises morphogenesis and the mechanical stability of
the embryo
The first three enzymatic activities responsible for
de novo pyrimidine biosynthesis in C elegans are
encoded by pyr-1 [29] The pyr-1()) mutants, like
umps-1()) mutants, exhibit partially penetrant
embry-onic lethality [31] (Table S2) The lethality of
pyr-1(cu8)is partially suppressed by the addition of uracil
[29], which can be converted into UMP via a salvage
pathway [30] Similarly, umps-1(zu456) viability was
substantially improved by the addition of uracil to the
growth medium (Table S3) Some of the lethality seen
in pyr-1()) mutants results from a pharyngeal
mor-phogenesis defect that leads to a pharynx-unattached
(Pun) phenotype The Pun phenotype is probably due
to loss of de novo formation of UMP that is utilized in
proteoglycan synthesis, which is known to be essential
for pharyngeal organogenesis [29] Like pyr-1())
embryos, umps-1()) embryos exhibited a partially
penetrant Pun phenotype (Table S1) The phenotypic
similarities between umps-1()) and pyr-1()) mutants,
together with the recent observation that umps-1(+)
activity is necessary for 5-fluorouracil-mediated toxicity
in C elegans [27], a process known to require a
func-tional pyrimidine biosynthesis pathway [31], and the
in vitro biochemical characterization of UMPS-1 [31],
indicate that C elegans UMPS-1 functions in de novo
pyrimidine biosynthesis
Embryonic gut granules are enlarged and not
properly formed in umps-1()) embryos
We investigated whether the vacuoles present in
umps-1()) embryos were enlarged gut granules The
umps-1()) vacuoles contained birefringent material,
and PGP-2 was localized to enlarged compartments in
umps-1()) embryos, suggesting that they were gut
granules (Fig 1J–L) The integral membrane gut granule-associated proteins PGP-2::green fluorescent protein (GFP) (data not shown) [15] and CDF-2::GFP [32] localized to the limiting membrane of the vacu-oles in umps-1()) embryos (Fig 3O,P) Comparison
of PGP-2-stained compartments in wild-type and umps-1()) pretzel-stage embryos showed average diameters of 0.41 ± 0.02 lm (n = 60) and 2.6 ± 0.05 lm (n = 50), respectively (± standard error of the mean) This represents a more than 250-fold increase in organelle volume in umps-1()) embryos
If the vacuoles in umps-1()) embryos are gut gran-ules, then their formation should depend on genes involved in the formation of gut granules Mutations disrupting the functions of the Rab GTPase GLO-1 [13], the AP-3 complex subunit APT-6 [13] and the ABC transporter PGP-2 [15] result in a Glo (gut gran-ule loss) phenotype We constructed umps-1()); glo()) double mutant embryos, and examined their intestinal cells for vacuoles The umps-1()); glo-1()) embryos completely lacked vacuoles, and umps-1()); apt-6()) embryos typically lacked vacu-oles (Table 1; Fig 4D,E) The umps-1()); pgp-2()) embryos exhibited small vacuoles containing birefrin-gent material (Table 1; Fig 4F), consistent with the partial defect in gut granule biogenesis seen in pgp-2()) embryos [15] We conclude that gut granules are enlarged in umps-1(zu456) embryos
The umps-1()) mutation affects the characteristics
as well as the size of gut granules Many gut granules
in umps-1()) embryos did not stain with Lysosensor Green DND-189 (Fig 3H), and none of them were stained by acridine orange (Fig 3D) Both of these markers of acidification accumulate in wild-type gut granules (Fig 3B,F) VHA-17, a subunit of the V-ATPase V0 domain [34], is present on gut granules and the apical surfaces of wild-type intestinal cells (Fig 3R) Although the apical localization was not altered, VHA-17-labeled compartments similar to those seen in wild type were lacking in umps-1()) embryos (Fig 5T) Detectable levels of VHA-17 were not asso-ciated with structures resembling enlarged gut granules (Fig 5T), consistent with the observed defects in gut granule acidification in umps-1()) embryos Unlike those in wild-type embryos (Fig 3J), gut granules in umps-1()) embryos did not stain with Nile Red (Fig 3L) These data demonstrate that the properties
of gut granules are dramatically altered in umps-1()) embryos At present, it is not clear whether this results from a defect in trafficking of material to the gut gran-ule or from a dilution of gut grangran-ule constituents due
to the dramatic enlargement of gut granule volume and surface area in umps-1()) embryos
Trang 7We tested whether the sizes of other endolysosomal
compartments were as dramatically altered as those of
gut granules in umps-1(zu456) embryos The
morphol-ogy of early endosomal-associated RAB-5::GFP [13]
and late endosomal-associated RAB-7::GFP [19] was
similar in umps-1(zu456) and wild-type embryos
(Fig 5B,D,F,H) RAB-5::GFP, RAB-7::GFP, and
the late endosome⁄ lysosome-associated LMP-1::GFP
proteins, which do not normally associate with gut
granules [15,33], were not obviously enriched on
the limiting membrane of umps-1()) vacuoles
(Fig 5C,D,G,H,L) Compartments containing
LMP-1::GFP [19] were slightly enlarged in umps-1()) embryos (Fig 5J,L) Additionally, LMP-1::GFP com-partments in umps-1()) 1.5-fold stage embryos were dispersed throughout the cytoplasm, and did not clus-ter near the apical surfaces of polarized intestinal cells,
as seen in wild-type embryos (Fig S3) It is possible that the altered cytoplasmic distribution of LMP-1::GFP-containing organelles is a consequence of extremely enlarged gut granules in umps-1()) embryos LMP-1::GFP is localized to lysosomal com-partments in C elegans phagocytic cells and coelomo-cytes [35,36] In C elegans embryonic intestinal cells,
Q
P
Fig 3 Gut granules are enlarged and their properties are altered in umps-1( )) pretzel-stage embryos In wild-type embryos, gut granules were acidified, being stained by acridine orange (B) and Lysosensor Green (F), contained lipid stained by Nile Red (J), and contained the inte-gral membrane proteins CDF-2::GFP (N) and VHA-17 (R) (gut granules are marked by white arrows in each panel) The vacuoles within umps-1( )) embryos did not accumulate acridine orange (D) or Nile Red (L); however, some vacuoles accumulated Lysosensor Green [white arrows in (H)] The umps-1(zu456) embryos contained greatly enlarged gut granules marked with CDF-2::GFP [white arrows in (P)] and lacked VHA-17-stained compartments within intestinal cells (T) The apical localization of VHA-17 was present in both wild-type and umps-1(zu456) embryos [black arrows in (R) and (T)] The intestine lies between the black arrowheads in all panels DAPI, 4¢,6-diamidino-2-phenylindole.
Trang 8we found that mCherry-tagged CPR-6 and F11E6.1
hydrolases were associated with
LMP-1::GFP-contain-ing organelles (Fig S3) The cpr-6 gene encodes a
cathepsin B protease, and F11E6.1 encodes a glucosyl-ceramidase, orthologs of which are found in mamma-lian conventional lysosomes [37] In umps-1(zu456)
A
B
C
D
E
F
Fig 4 Suppression of umps-1( )) vacuole
formation DIC microscopy was used to
ana-lyze embryos for intestinal vacuoles, which
are prominent in umps-1( )) embryos [white
arrows in (A)] The umps-1( )); pyr-1()) (B)
and umps-1( )); wht-2()) (C) embryos lacked
vacuoles and elongated normally The
umps-1( )); apt-6()) (D) and umps-1());
glo-1(zu437) (E) embryos lacked vacuoles
and did not elongate beyond the 1.25-fold
stage The umps-1( )); pgp-2()) embryos
contained small vacuoles [white arrow in (F)]
and arrested elongation prior to the 1.5-fold
stage The umps-1( )) embryos display
vacuoles from the bean stage through
embryogenesis (Fig S2) White arrowheads
(A, B) flank the pharynx of an embryo
exhibiting the Pun phenotype Black
arrow-heads flank the intestine in all panels.
Fig 5 Analysis of endosomal compartments in umps-1( )) embryos The size and morphology of RAB-5::GFP-labeled endosomes [white arrows in (B) and (D)] and RAB-7::GFP-labeled endosomes [white arrows in (F) and (H)] were similar in wild-type and umps-1( )) pretzel-stage embryos LMP-1::GFP-containing compartments were slightly enlarged in umps-1( )) embryos [compare white arrows in (J) and (L)] Black arrowheads flank the intestine in all panels.
Trang 9embryos, both proteins were localized to
LMP-1::GFP-labeled compartments, suggesting that these
organelles are properly formed in umps-1(zu456)
embryos (Fig S3) Thus, umps-1()) appears to most
dramatically affect the formation and morphology of
gut granules
A role for the ABC transporter WHT-2 in
umps-1()) gut granule enlargement
Lysosomal compartments are highly sensitive to
osmo-tic stress, showing rapid vacuolization on the
accumu-lation of osmotically active material within the
lysosomal lumen [38,39] Therefore, material within the
gut granule could have a significant impact on its size
Gut granules contain birefringent material [13,14],
cur-rently of unknown composition [33] As the
birefrin-gent material is probably present at a high
concentration within the gut granule, we examined its
role in the vacuolization of umps-1()) gut granules
Disrupting the function of the ABC transporters
MRP-4 and WHT-2 delays the appearance of
birefrin-gent material within gut granules, but does not
other-wise obviously disrupt gut granule biogenesis [33]
(data not shown) The mrp-4()); umps-1()) double
mutant embryos displayed normal-sized vacuoles,
many of which lacked birefringent material, indicating
that the formation of birefringent granules per se is
not required for gut granule enlargement (Table 1)
We used both wht-2(RNAi) and a wht-2 deletion
allele, wht-2(ok2775), to disrupt wht-2(+) activity In
all of the wht-2()); umps-1()) double mutant
combi-nations, we examined whether there was a loss of, or a
significant reduction in, the number of vacuoles within
intestinal cells (Fig 4C; Table 1) The wht-2(ok2775)
allele also partially suppressed the embryonic lethality
of umps-1(zu456) (Table S2) Anti-PGP-2 staining
showed that gut granule size was reduced from
an average diameter of 2.6 ± 0.05 lm (n = 50) in
umps-1(zu456) embryos to 0.66 ± 0.04 lm (n = 51)
in umps-1(zu456); wht-2(ok2775) double mutants
(Fig 6E) In addition, the gut granules in
umps-1(zu456); wht-2(ok2775) embryos were stained by
acridine orange (data not shown) Forty other ABC
transporter mutants were unable to suppress the
for-mation of vacuoles in umps-1(RNAi) embryos
(Table S4) These results indicate that wht-2(+)
activ-ity is necessary for the enlargement of gut granules in
umps-1()) embryos The lack of similar suppression
by mrp-4()) suggests that wht-2()) mediates this
effect through processes independent of the
accumula-tion of birefringent material within gut granules
We noticed that many pretzel-stage umps-1()); wht-2()) double mutant embryos exhibited a Pun phenotype Nearly 50% of umps-1(zu456); wht-2 (ok2775) and umps-1(zu456); wht-2(RNAi) embryos exhibited a Pun phenotype, whereas wht-2()) embryos did not, and only 7% of umps-1(zu456) embryos did (Table S1) We investigated whether the genetic inter-action leading to the Pun phenotype was between wht-2()) and the de novo pyrimidine biosynthetic pathway or was specific to umps-1()) pyr-1(cu8); wht-2(RNAi) embryos did not exhibit a Pun pheno-type (Table S1), indicating that the Pun phenopheno-type of umps-1()); wht-2()) represents a specific genetic interaction between these two genes The genetic inter-actions between umps-1()) and wht-2()) implicate the WHT-2 ABC transporter in the trafficking of metabo-lites that accumulate in umps-1()) embryos, which ultimately impinge upon gut granule size and pharyn-geal morphogenesis
Analysis of UMPS-1 expression, localization, and function
To investigate where UMPS-1 functions and how it might directly regulate gut granule morphology, we expressed a umps-1::gfp gene under control of the 2.7 kb umps-1 promoter The UMPS-1::GFP fusion rescued the Vac phenotype of umps-1(zu456) embryos (Table 1) UMPS-1::GFP expression was first detected
in early pretzel-stage embryos, where it was expressed
in the intestine and in a few cells in the head and tail
of the animal (Fig 7A,B) In larval (not shown) and adult stages, UMPS-1::GFP was expressed in the intes-tine and neuronal cells located near the nerve ring and rectum (Fig 7E,F), which is similar to what has been documented for an umps-1 promoter-driven reporter [27]
The umps-1(zu456) embryos displayed a strict, maternal effect Vac phenotype (Table 1) This could result from metabolic processes involving UMPS-1 at work in the adult intestine that impact on embryonic gut granules For example, yolk proteins derived from the adult intestine are transferred into oocytes, where they accumulate in the embryonic intestine [40,41]
We performed RNAi on rde-1()); elt-2p::rde-1(+) animals, which are only susceptible to feeding based RNAi in larval and adult intestinal cells [42]
We found that none of the embryos exhibited a Vac phenotype (Table 1), suggesting that inhibiting umps-1(+) in the adult intestine does not impact on embryonic gut granule size
We next considered whether the loss of umps-1 expression in the germline leads to enlarged gut
Trang 10granules We analyzed the effects of umps-1(RNAi)
on rrf-1(pk1417) animals, which are defective for somatic RNAi but are competent for germline RNAi [43] We found that rrf-1(pk1417) animals were as sensitive to umps-1(RNAi) as wild-type animals (Table 1) Thus, umps-1 expression in the germline is necessary to prevent the enlargement of embryonic gut granules, suggesting that the maternal effect Vac phenotype of umps-1()) probably results from the maternal contribution of umps-1(+) to embryonic progeny This could take the form of UMPS-1 protein, umps-1 mRNA, and⁄ or UMPS-1 metabolic activity in the germline
Mammalian UMPS is localized to the cytoplasm [44,45], and C elegans UMPS-1 does not contain any obvious organelle targeting or retention motifs, sug-gesting a similar localization In embryonic intestinal cells, UMPS-1::GFP was distributed throughout the cytoplasm, without any obvious organelle association (Fig 7A,B) However, we often observed UMPS-1::GFP near the apical surface of the embryonic intestine (Fig 7A,B) In adult intestinal cells, UMPS-1::GFP appeared to be uniformly localized throughout the cytoplasm (Fig 7C,D) These data suggest that UMPS-1 is a cytoplasmic protein that is not associated with the gut granule
Accumulation of orotic acid probably leads to enlarged gut granules in umps-1()) embryos Mutations that disrupt the function of human UMPS result in orotic aciduria, a disease characterized by megaloblastic anemia, failure to thrive, and urinary excretion of large amounts of orotic acid [6,30,46] Disrupting the function of the Drosophila UMPS-encoding gene rudimentary-like results in sterility, reduced viability, wing and leg morphological defects, and accumulation of orotic acid [47–49]
Many of the phenotypes resulting from loss of UMPS activity are due to pyrimidine auxotrophy [30,47] We therefore considered the possibility that
A
B
C
D
E
Fig 6 Suppression of enlarged gut granules in umps-1( )) embryos Embryos lacking umps-1(+) activity had enlarged gut granules marked with antibodies against PGP-2 [white arrows in (B)] The pyr-1( )) embryos had gut granules that were slightly enlarged [white arrows in (C)] in comparison with wild-type embryos [white arrow in (A)] The gut granules of umps-1( )); pyr-1( )) and umps-1()); wht-2()) embryos were dramatically reduced in size [white arrows in (D) and (E)], and were similar in size to gut granules in pyr-1( )) embryos [white arrows in (C)] The intestine of pretzel-stage embryos is flanked by black arrowheads
in all panels.