Developmental control of histone mRNA and dSLBP synthesis during Drosophila embryogenesis and the role of dSLBP in histone mRNA 3' end processing in vivo

Abstract

In metazoans, the 3' end of histone mRNA is not polyadenylated but instead ends with a stem-loop structure that is required for cell cycle-regulated expression. The sequence of the stem-loop in the Drosophila melanogaster histone H2b, H3, and H4 genes is identical to the consensus sequence of other metazoan histone mRNAs, but the sequence of the stem-loop in the D. melanogaster histone H2a and H1 genes is novel. dSLBP binds to these novel stem-loop sequences as well as the canonical stem-loop with similar affinity. Eggs derived from females containing a viable, hypomorphic mutation in dSLBP store greatly reduced amounts of all five histone mRNAs in the egg, indicating that dSLBP is required in the maternal germ line for production of each histone mRNA. Embryos deficient in zygotic dSLBP function accumulate poly(A)(+) versions of all five histone mRNAs as a result of usage of polyadenylation signals located 3' of the stem-loop in each histone gene. Since the 3' ends of adjacent histone genes are close together, these polyadenylation signals may ensure the termination of transcription in order to prevent read-through into the next gene, which could possibly disrupt transcription or produce antisense histone mRNA that might trigger RNA interference. During early wild-type embryogenesis, ubiquitous zygotic histone gene transcription is activated at the end of the syncytial nuclear cycles during S phase of cycle 14, silenced during the subsequent G(2) phase, and then reactivated near the end of that G(2) phase in the well-described mitotic domain pattern. There is little or no dSLBP protein provided maternally in wild-type embryos, and zygotic expression of dSLBP is immediately required to process newly made histone pre-mRNA.

FIG. 1.

dSLBP binds to all five histone mRNAs. (A) The sequence of the 3′ end of the most common stem-loop sequence in vertebrate histone mRNAs (H2a_vert) and the sequence of the 3′ end of the five Drosophila histone mRNAs is shown. ∗, nucleotide pairs that are different than the canonical sequence and that have not yet been observed in any other stem-loop sequence. (B) DNAs encoding dSLBP (lanes 1 and 2), xSLBP1 (lanes 3 and 4), and xSLBP2 (lanes 5 and 6) were incubated in a reticulocyte lysate-coupled transcription-translation system in duplicate in the presence of [³⁵S]methionine. The products were analyzed on an SDS-10% polyacrylamide gel and detected by autoradiography. Lane 7, lysate incubated without added DNA. Parallel reaction mixtures incubated in the absence of [³⁵S]methionine were used for the mobility shift assays (C and D). (C) The indicated stem-loop sequences (lanes 1 to 3, H2a_vert; lanes 4 to 6, Drosophila H2a; lanes 7 to 9, Drosophila H1; lanes 10 and 11, Drosophila H2b) labeled with ³²PO₄ were tested for their ability to bind to Xenopus SLBP1 (X1 [lanes 1, 4, 7, and 11]), Xenopus SLBP2 (X2 [lanes 2, 5, and 8]), or Drosophila SLBP (D [lanes 3, 6, and 9]). Lane 12, probe incubated in buffer. (D) xSLBP1, xSLBP2, and dSLBP (indicated above each panel) were incubated with the H2a_vert probe (lanes 2 to 5), dH2a probe (lanes 6 to 8), or dH1 probe (lanes 9 to 11), and the complexes were detected by mobility shift assay. Increasing amounts of unlabeled H2a_vert RNA (indicated above each lane) were mixed with the probe prior to addition of the SLBPs. Lane 1, probe incubated in buffer. The complexes were analyzed by electrophoresis on an 8% polyacrylamide gel.

FIG. 1.

dSLBP binds to all five histone mRNAs. (A) The sequence of the 3′ end of the most common stem-loop sequence in vertebrate histone mRNAs (H2a_vert) and the sequence of the 3′ end of the five Drosophila histone mRNAs is shown. ∗, nucleotide pairs that are different than the canonical sequence and that have not yet been observed in any other stem-loop sequence. (B) DNAs encoding dSLBP (lanes 1 and 2), xSLBP1 (lanes 3 and 4), and xSLBP2 (lanes 5 and 6) were incubated in a reticulocyte lysate-coupled transcription-translation system in duplicate in the presence of [³⁵S]methionine. The products were analyzed on an SDS-10% polyacrylamide gel and detected by autoradiography. Lane 7, lysate incubated without added DNA. Parallel reaction mixtures incubated in the absence of [³⁵S]methionine were used for the mobility shift assays (C and D). (C) The indicated stem-loop sequences (lanes 1 to 3, H2a_vert; lanes 4 to 6, Drosophila H2a; lanes 7 to 9, Drosophila H1; lanes 10 and 11, Drosophila H2b) labeled with ³²PO₄ were tested for their ability to bind to Xenopus SLBP1 (X1 [lanes 1, 4, 7, and 11]), Xenopus SLBP2 (X2 [lanes 2, 5, and 8]), or Drosophila SLBP (D [lanes 3, 6, and 9]). Lane 12, probe incubated in buffer. (D) xSLBP1, xSLBP2, and dSLBP (indicated above each panel) were incubated with the H2a_vert probe (lanes 2 to 5), dH2a probe (lanes 6 to 8), or dH1 probe (lanes 9 to 11), and the complexes were detected by mobility shift assay. Increasing amounts of unlabeled H2a_vert RNA (indicated above each lane) were mixed with the probe prior to addition of the SLBPs. Lane 1, probe incubated in buffer. The complexes were analyzed by electrophoresis on an 8% polyacrylamide gel.

FIG. 1.

dSLBP binds to all five histone mRNAs. (A) The sequence of the 3′ end of the most common stem-loop sequence in vertebrate histone mRNAs (H2a_vert) and the sequence of the 3′ end of the five Drosophila histone mRNAs is shown. ∗, nucleotide pairs that are different than the canonical sequence and that have not yet been observed in any other stem-loop sequence. (B) DNAs encoding dSLBP (lanes 1 and 2), xSLBP1 (lanes 3 and 4), and xSLBP2 (lanes 5 and 6) were incubated in a reticulocyte lysate-coupled transcription-translation system in duplicate in the presence of [³⁵S]methionine. The products were analyzed on an SDS-10% polyacrylamide gel and detected by autoradiography. Lane 7, lysate incubated without added DNA. Parallel reaction mixtures incubated in the absence of [³⁵S]methionine were used for the mobility shift assays (C and D). (C) The indicated stem-loop sequences (lanes 1 to 3, H2a_vert; lanes 4 to 6, Drosophila H2a; lanes 7 to 9, Drosophila H1; lanes 10 and 11, Drosophila H2b) labeled with ³²PO₄ were tested for their ability to bind to Xenopus SLBP1 (X1 [lanes 1, 4, 7, and 11]), Xenopus SLBP2 (X2 [lanes 2, 5, and 8]), or Drosophila SLBP (D [lanes 3, 6, and 9]). Lane 12, probe incubated in buffer. (D) xSLBP1, xSLBP2, and dSLBP (indicated above each panel) were incubated with the H2a_vert probe (lanes 2 to 5), dH2a probe (lanes 6 to 8), or dH1 probe (lanes 9 to 11), and the complexes were detected by mobility shift assay. Increasing amounts of unlabeled H2a_vert RNA (indicated above each lane) were mixed with the probe prior to addition of the SLBPs. Lane 1, probe incubated in buffer. The complexes were analyzed by electrophoresis on an 8% polyacrylamide gel.

FIG. 1.

dSLBP binds to all five histone mRNAs. (A) The sequence of the 3′ end of the most common stem-loop sequence in vertebrate histone mRNAs (H2a_vert) and the sequence of the 3′ end of the five Drosophila histone mRNAs is shown. ∗, nucleotide pairs that are different than the canonical sequence and that have not yet been observed in any other stem-loop sequence. (B) DNAs encoding dSLBP (lanes 1 and 2), xSLBP1 (lanes 3 and 4), and xSLBP2 (lanes 5 and 6) were incubated in a reticulocyte lysate-coupled transcription-translation system in duplicate in the presence of [³⁵S]methionine. The products were analyzed on an SDS-10% polyacrylamide gel and detected by autoradiography. Lane 7, lysate incubated without added DNA. Parallel reaction mixtures incubated in the absence of [³⁵S]methionine were used for the mobility shift assays (C and D). (C) The indicated stem-loop sequences (lanes 1 to 3, H2a_vert; lanes 4 to 6, Drosophila H2a; lanes 7 to 9, Drosophila H1; lanes 10 and 11, Drosophila H2b) labeled with ³²PO₄ were tested for their ability to bind to Xenopus SLBP1 (X1 [lanes 1, 4, 7, and 11]), Xenopus SLBP2 (X2 [lanes 2, 5, and 8]), or Drosophila SLBP (D [lanes 3, 6, and 9]). Lane 12, probe incubated in buffer. (D) xSLBP1, xSLBP2, and dSLBP (indicated above each panel) were incubated with the H2a_vert probe (lanes 2 to 5), dH2a probe (lanes 6 to 8), or dH1 probe (lanes 9 to 11), and the complexes were detected by mobility shift assay. Increasing amounts of unlabeled H2a_vert RNA (indicated above each lane) were mixed with the probe prior to addition of the SLBPs. Lane 1, probe incubated in buffer. The complexes were analyzed by electrophoresis on an 8% polyacrylamide gel.

FIG. 2.

The dSLBP¹⁰ maternal effect lethal allele causes reduced deposition of maternal histone mRNA. (A) Total RNA extracted from 0- to 2-h-old embryos laid by wild-type or dSLBP¹⁰/Df(3R)3450 mutant mothers were subjected to Northern analysis using radiolabeled probes for H2a, H2b, and H1. The blots were stripped and reprobed with rp49 as a loading control. (B) Detection of H2a mRNA by S1 nuclease protection. A 3′ end-labeled 650-nt H2a probe (lane 1) was incubated with S1-containing buffer (lane 2), nonspecific yeast tRNA (lane 3), a synthetic partial histone H2a mRNA that should protect a 265-nt fragment (lane 4), or total RNA from 0- to 2-h-old wild-type (lane 5) or dSLBP¹⁰/Df(3R)3450 embryos (lane 6). The S1 nuclease-resistant fragments were resolved by gel electrophoresis. In both samples, a fragment of 340 nt is observed which corresponds to histone H2a mRNA ending at the normal processing site (arrow). The upper 650-nt fragment is undigested probe (P). A diagram of the H2a S1 assay is shown on the bottom.

FIG. 3.

Altered histone mRNA processing in dSLBP mutant embryos. Total RNA isolated from 13- to 16-h-old wild-type or hand-selected GFP-negative mutant embryos was subjected to Northern analysis using a radiolabeled H4, H3, H2b, or H2a probe. Lane 1, yw⁶⁷ wild-type; lane 2, dSLBP¹⁵/TM3 [actin-GFP] males mated to Df(3R)3450/TM3 [actin-GFP] females; lane 3, dSLBP¹⁰/TM3 [actin-GFP] males mated to Df(3R)3450/TM3 [actin-GFP] females; lane 4, Df(3R)3450/TM3 [actin-GFP] males mated to dSLBP¹⁵/TM3 [actin-GFP] females (the reciprocal cross to lane 2); lanes 5 and 6, GFP-positive embryos (two of three of which are dSLBP/+, and one of three are +/+) selected from the same collections used for lane 2 and lane 4, respectively. Each blot was stripped and reprobed with a radiolabeled rp49 probe as a loading control.

FIG. 4.

S1 nuclease protection analysis of histone mRNA in dSLBP mutant embryos. (A) Total RNA isolated from 13- to 16-h-old wild-type or hand-selected GFP-negative mutant embryos was subjected to S1 nuclease protection analysis using a ³²P end-labeled histone H2aprobe as described for Fig. 2B. Lane 1, in vitro-transcribed H2a RNA as a positive control, designed to yield a 265-nt protection product; lane 2, yw⁶⁷ wild type. The arrow indicates the 340-nt H2a-protected fragment corresponding to the normally processed form, and the upper band is undigested probe (P) (Fig. 2). Lane 3, dSLBP¹⁵/TM3 [actin-GFP] males mated to Df(3R)3450/TM3 [actin-GFP] females; lane 4, dSLBP¹⁰/TM3 [actin-GFP] males mated to Df(3R)3450/TM3 [actin-GFP] females; lane 5, Df(3R)3450/TM3 [actin-GFP] males mated to dSLBP¹⁵/TM3 [actin-GFP] females (the reciprocal cross to lane 3); lane 6, GFP-positive embryos (two of three of which are dSLBP/+ and one of three of which are +/+) selected from the same collections used for lane 2. Multiple protected fragments corresponding to misprocessed forms of H2a mRNA are observed. The major misprocessed forms are indicated (∗ and ∗∗). A diagram of the S1 nuclease assay is shown below the gel. (B) S1 analysis using a ³²P end-labeled probe designed to detect histone H1 mRNA. Lane 1, H1 probe plus S1 nuclease; lane 2, H1 probe plus a positive control synthetic H1 RNA that generates a predicted fragment of 260 nt; lane 3, yw⁶⁷ wild-type RNA; lane 4, dSLBP¹⁵/TM3 [actin-GFP] males mated to Df(3R)3450/TM3 [actin-GFP] females; lane 5, dSLBP ¹⁰/TM3 [actin-GFP] males mated to Df(3R)3450/TM3 [actin-GFP] females; lane 6, Df(3R)3450/TM3 [actin-GFP] males mated to dSLBP¹⁵/TM3 [actin-GFP] females. A diagram of the S1 nuclease assay is shown below the gel. Only the normally processed form of H1 mRNA is observed in the wild-type sample at the expected size of 560 nt (arrow). Mutant extracts contain an additional fragment corresponding to misprocessed H1 mRNA at 650 nt (asterisk).

FIG. 5.

RT-PCR mapping of the location of cryptic polyadenylation sites in misprocessed histone messages from dSLBP mutant embryos. (A) Diagram of the 5-kb repeat unit of the Drosophila histone gene cluster. Arrows indicate the direction and length of each histone transcript, which do not contain introns. (B to D) Poly(A)⁺ mutant histone mRNA was amplified by RT-PCR using an oligo(dT) primer. Cloned products were randomly selected and sequenced to identify the location of polyadenylation. These sites of polyadenylation are indicated by a hatchmark (|) along with the number of times that the particular location was observed. RT-PCR products specific to the histone proceeding (5′ → 3′) top-to-bottom are shown with the number of products above the hatchmark. RT-PCR products specific to the histone proceeding (5′ → 3′) bottom-to-top are shown below the hatchmark. For each gene, the translation termination codon is shown in bold, the processing site is indicated (::), and the stem-loop sequence is depicted in bold and double underlined. Canonical polyadenylation signal sequences downstream of the processing sites are shown in bold with a single underline. Those signal sequences properly spaced with respect to polyadenylation sites determined by RT-PCR products are shown in italics. RT-PCR products corresponding in size to S1 products in Fig. 4 are shown with matching asterisks (∗, ∗∗). (B) H4/H2a intergenic sequence; (C) H2b/H1 intergenic sequence; (D) The downstream sequence of H3. RT-PCR products mapping to putative polyadenylation signal sequences are observed for histones H4, H2a, H2b, and H1. 3′-End mapping of misprocessed histone H3 mRNA was hampered by the presence of oligo(A) (boxed) sequences in the template mRNA. The DNA sequence included in the H3-ds probe used for detection of misprocessed histone H3 mRNA by embryonic in situ hybridization in Fig. 7 and 9 is indicated with brackets.

FIG. 6.

Altered histone mRNA expression in dSLBP mutant embryos. Each panel shows a whole mount in situ hybridization of a stage 14 embryo (anterior at left, dorsal at top) using a coding region histone probe. (A) Wild-type H2a expression is restricted to replicating cells, including those in the proliferating CNS (arrowhead) and endoreduplicating midgut (right arrow). The cells of the anterior midgut (left arrow) have exited S phase and do not express histone H2a mRNA. (B) Wild-type histone H2b expression is identical to H2a. (C) In dSLBP¹⁵/dSLBP¹⁵ mutant embryos, histone H2a mRNA persists in cells that have finished replicating (e.g. the anterior midgut; arrow), generating a diagnostic aberrant pattern. (D) H2b mRNA also accumulates in nonreplicating dSLBP¹⁵/dSLBP¹⁵ mutants cells (arrow) in a manner similar to H2a mRNA.

FIG. 7.

In situ hybridization with a downstream probe complementary to the 3′ untranscribed region of histone H3 mRNA to detect misprocessed H3 mRNA in vivo. The sequence used to generate the antisense H3-ds probe for use in detecting only misprocessed forms of histone H3 is shown by brackets in Fig. 5D. Each panel shows a whole embryo oriented as in Fig. 6 that was hybridized with the H3-ds probe (A to D, F, and H) or a coding region H3 probe. (E and G). The embryos stained with the H3-ds probe are siblings collected from heterozygous dSLBP¹⁵/+ parents and provide examples of the three different phenotypic classes observed. (A) Stage 12 +/+. Most cells are beginning to enter cell cycle 17 at this time. (B) Stage 12 dSLBP¹⁵/+; (C) stage 13 +/+; (D) stage 13 dSLBP¹⁵/+; (E) stage 12 yw⁶⁷ wild type; (F) stage 12 dSLBP¹⁵/dSLBP¹⁵; (G) stage 14 yw⁶⁷ wild type; (H) stage 14 dSLBP¹⁵/dSLBP¹⁵. Persistence of misprocessed histone H3 message is observed in nonreplicating cells of the pharanx, anterior midgut, hindgut, and anal pads (indicated from left to right, respectively, by arrows in H and G).

FIG. 8.

dSLBP function is required as soon as zygotic transcription begins. Appearance of misprocessed histone mRNA in dSLBP mutant embryos is coincident with the normal onset of dSLBP protein expression. (A) Total RNA from 13- to 16-h-old wild-type embryos (lane 1) and the poly(A)⁺ RNA fraction from appropriately aged embryos collected from dSLBP¹⁵/+ parents corresponding to 0 to 2 h (lane 2), 2 to 4 h (lane 3), 4 to 7 h (lane 4), 7 to 10 h (lane 5), 10 to 13 h (lane 6), and 13 to 16 h (lane 7) AED was subjected to Northern analysis using a radiolabeled H3 coding region probe. Polyadenylated H3 mRNA begins to accumulate at ∼2 h of embryogenesis, which is the time when most zygotic gene expression begins. The blot was stripped and reprobed for rp49 as a loading control. (B) Total RNA isolated from wild-type embryos aged 0 to 1 h (lane 1), 1 to 2 h (lane 2), 2 to 4 h (lane 3), 4 to 7 h (lane 4), 7 to 10 h (lane 5), 10 to 13 h (lane 6), and 13 to 16 h (lane 7) AED was probed with the histone H3 coding region. (C and D) Western analysis of embryonic extracts using affinity-purified anti-dSLBP rabbit antibodies. dSLBP protein is indicated with an asterisk. All other bands represent nonspecific cross-reactingproteins (48). (C) Lane 1, 0- to 2-h-old embryos collected from dSLBP¹⁰/Df(3R)3450 females; lane 2, wild-type 13- to 16-h-old embryos; lane 3, wild-type 13- to 16-h-old embryo extract treated with CIP. Note that dSLBP migrates faster after phosphatase treatment. Lane 4, 13- to 16-h-old dSLBP¹⁵/Df(3R)3450 embryos. Extracts of wild-type yw⁶⁷ embryos collected 0 to 1 h (lane 5), 1 to 2 h (lane 6), 2 to 4 h (lane 7) AED. (D) Extracts of wild-type yw⁶⁷ embryos collected 0 to 1 h (lane 1), 1 to 2 h (lane 2), 2 to 4 h (lane 3), 4 to 7 h (lane 4), 7 to 10 h (lane 5), 10 to 13 h (lane 6), and 13 to 16 h (lane 7) AED. These are the same embryo collections used to prepare the RNA for panel B.

FIG. 9.

Zygotic histone transcription occurs in the mitotic domain pattern in the early embryo. (A to D) In situ hybridization of yw⁶⁷ wild-type embryos using an H3 coding region probe. (A) A high level of maternal histone H3 mRNA is deposited into the egg. (B) Much lower levels are observed in cellular blastoderm embryos. The detectable H3 message may result from a failure to destroy all maternal H3 mRNA and/or from the initiation of zygotic histone transcription during S phase 14. (C) As embryos begin to gastrulate in cycle 14, zygotic H3 mRNA begins to accumulate in the mitotic domain pattern. The arrows indicate mitotic domains 1, 2, 5, and 6. The mitotic domain pattern is somewhat masked by the presence of ubiquitous H3 mRNA. (D) H3 mRNA accumulation becomes widespread in germ band extended embryos, although the mitotic domain pattern is still evident. (E to I) In situ hybridization of embryos collected from dSLBP¹⁵/+ parents using the histone H3-ds probe. (E) No misprocessed maternal H3 mRNA is detected in either pre-blastoderm or blastoderm embryos regardless of genotype. (F) During gastrulation de novo synthesis of misprocessed histone H3 mRNA in dSLBP¹⁵/dSLBP¹⁵ embryos begins to appear in the mitotic domain pattern, with domains 1, 2, 5, and 6 indicated by arrows. This embryo is slightly older than the embryo shown in panel C. (G) Higher magnification of misprocessed histone H3expression reveals nascent transcription dots (arrow) in the nucleus of cells from dSLBP¹⁵/dSLBP¹⁵ mutant embryos. Mitotic domain 2 is shown. (H) Stage 11 (cycle 17) dSLBP¹⁵/dSLBP¹⁵ embryo showing that misprocessed H3 mRNA eventually accumulates in the cytoplasm. Nascent transcription dots can also be seen at this stage (arrow). (I) A cellularizing blastoderm embryo where nascent transcription dots are observed during S14. A single transcription dot is visible in some nuclei, while others have two (arrow). One versus two likely depends on the degree of pairing between homologs and/or the orientation of the nucleus with repect to the focal plane (28).

FIG. 10.

Model of histone gene expression in the early Drosophila embryo. Maternal histone mRNA is rapidly destroyed at interphase 14. Zygotic histone transcription occurs during S phase of cycle 14 as determined by transcription dots visible with the H3-ds 3′ flanking probe and terminates by the start of G₂. Transcription reinitiates in late G₂ of cycle 14 in each mitotic domain in anticipation of S phase of cycle 15. A low level of histone mRNA can be detected in all cells during G₂ of cycle 14, probably because both the maternal mRNA and mRNA synthesized during S phase 14 are not quantitatively destroyed at the end of S phase 14. dSLBP protein is not provided maternally, and the timing of its accumulation zygotically matches the onset of histone transcription in S phase 14. When zygotic dSLBP is not provided, misprocessed histone messages begin to accumulate immediately in S phase 14 and subsequently in G₂ phase 14.