A) Molecular map of the abdominal region of the Bithorax complex numbered in kb according to (Genbank U31961). The abd-A and Abd-B transcription units are drawn below the DNA line along with the extent of the segment-specific iab cis-regulatory domains iab-2 through iab-9. B) The rectangle depicts the extent of the BAC used in this study. Note that it lacks the B,C and γ promoters specific for the Abd-Br form. The Gal-4 coding sequence was inserted within the 5′UTR of the Abd-Bm form. C) The structure of the vector sequences used to propagate the BAC and to select the integration within the Drosophila genome. Note the presence of two gypsy insulator sequences flanking the mini-white sequences to prevent possible position effect on white expression (see for further details).

A) Embryo expressing the Abd-B-Gal4 BAC crossed to a UAS-LacZ reporter stained with an antibody directed against ß-galactosidase. Out of a slight ectopic expression anteriorly (indicated by the arrow), the expression pattern is stickingly simiar to the WT Abd-B expression pattern as documented in . B) Cartoon depicting the male reproductive apparatus with testis, the paired accessory glands, the ejaculatory duct and ejaculatory bulb. Each accessory gland contains two secretory cell types, the main cells which make up the majority of the gland (top insert) and the secondary cells which are located at the distal tip of the gland interspersed among the main cells (bottom insert) Drawing by J. L. Sitnik; C) Picture of the male reproductive system from flies carrying the Abd-B-Gal4 BAC crossed to a UAS-GFP reporter with the secondary cells of the accessory glands showing GFP expression. The different organs composing the system are marked. D) Magnification of three secondary cells from flies carrying the Abd-B-Gal4 BAC crossed to a cytoplasmic UAS-GFP reporter. The multiple, large vacuoles, characteristic of secondary cells, can be visualized through their exclusion of the GFP protein. The two nuclei of the cells can also be seen as slightly more intense GFP signals; E) The tip of the accessory gland with GFP expressed specifically in the secondary cells driven by the Abd-B-Gal4 BAC (green), co-stained with the membrane staining dye, FM4–64, in red. The two cell types can be clearly distinguished with examples indicated with white lines.; F) Abd-B antibody staining of the tip of an accessory gland on Abd-B-Gal4 BAC, UAS-GFP flies (red). Only secondary cell nuclei are stained. G) GFP expression (green) in the same gland overlaid onto the Abd-B antibody staining (red) shown in . Each cell with Abd-B protein expression also express GFP. The white scale bar on figures D, F, E, and G represents 50 µm.

A) The Molecular map of the Abd-B gene region is shown with its extensive 3′ cis-regulatory domains iab-5 through iab-8 (the iab-4 domain regulates abd-A). The extents of the various deficiencies that were used to map the enhancer responsible for Abd-B expression in the secondary cells are shown below the molecular map. The location of DNA sequence used to make the 2.8 kb-long D5rsG4rs driver (thereby refereed as D5 Gal4 driver).is shown under the map. The red circles on the map represent the boundaries separating the parasegment-specific cis-regulatory domains of Abd-B. The green triangle above the iab-6 domain marks the iab-6 initiator. B) UAS-GFP expression driven by Abd-B-Gal4 in a WT for the BX-C. C) same as B, but in an iab-6, 7^IH homozygous male or in an iab-5,6^J82 homozygous male(D). Note that in the iab-6, 7^IH and iab-5,6^J82 background, the numerous vacuoles, characteristic of the secondary cells (visible by black holes in the GFP background), are lost. However, the vacuoles are not affected in iab-4,5,6^DB. Thus, the critical region required for proper secondary cell specification based on these 3 deficiencies is indicated by the dotted-line box in panel A. E) UAS-GFP expression driven by Abd-B-Gal4 in secondary cells of iab-6⁴ (initiator deletion) and of iab-6^cocu males (F). Note the normal aspect of GFP staining in iab-6⁴ (E) relative to the WT shown in B). In iab-6^cocu however (F), the vacuoles are lost, giving rise to staining comparable to panels C and D. Panels G) and H) show iab-6⁴ (G) and iab-6^cocu (H) accessory glands stained with an Abd-B antibody. While Abd-B expression appears normal in iab-6⁴ (G), the signal is absent in iab-6^cocu (H). I) shows the tip of an accessory gland from a fly carrying the D5-Gal4 driver driving GFP expression in the secondary cells (the staining is shown in yellow to distinguish it from panels B–F depicting GFP driven by the Gal4 Bac. The white horizontal scale bars in each of the panels represents 50 µm.

A) Western blots of accessory gland extracts from two control males (lane 1), two iab-6^cocu males (lane 2), two wild type males (lane 3), and two DTA-E males (lane 4). All Acps known to be produced by the primary cell (Acp36DE, Acp62F, sex peptide, and ovulin) are present in the accessory glands of iab-6^cocu males, but not DTA-E males. B) Other Acps necessary for various aspects of the PMR (CG11864, Seminase, CG1656, CG1652, CG17575, and CG9997) are present in the accessory glands of iab-6^cocu males. CG1656, CG1652, and CG17575 are always detectable in DTA-E males, however their abundance is highly variable compared to controls (The western blots depicted were selectived to most clearly demonstrate the presence of these proteins in DTA-E males). C) Mates of iab-6^cocu males have less SP, as detected by antibodies to SP, in the reproductive tract at 1 d ASM and all subsequent time points. Tubulin was used as a loading control for the female reproductive tracts. Accessory gland extracts from a single control male (lane 1) and iab-6^cocu male (lane 2) were used as a positive control and accessory gland extracts from 2 DTA-E males (lane 3) and reproductive tract extracts from 8 virgin females (lane 4) were used as a negative control. Reproductive tract extracts from females mated to either control (+) or iab-6^cocu (Δ) males at 2 h (lane 5–6, 2 RTs per), 1 d (lane 7–8, 20 RTs per), 4 d (lane 9–10, 18 RTs per), and 7 d ASM (lane 11–12, 21 RTs per). D) Mates of iab-6^cocu males have dramatically less SP in the seminal receptacle (SR) at 2 h ASM. Tubulin was used as a loading control for the female reproductive tracts. Accessory gland extracts from a single control male (lane 1) and iab-6^cocu male (lane 2) were used as positive controls and reproductive tract extracts from 8 virgin females (lane 3) were used as a negative control. Extracts from SRs dissected from females mated to either control (+) or iab-6^cocu (Δ) males at 2 h ASM (20 SRs each, lanes 4–5). The amount of SP present in the reproductive tract (minus the SR) of mates of control and iab-6^cocu males was determined in a dilution series (1∶1 (lanes 6 and 9), 1∶2 (lanes 7 and 10), and 1∶4 (lanes 8 and 11) and are equivalent to 5 RTs, 2.5 RTs, and 1.25 RTs). There is no appreciable difference in the amount of transferred SP.

A) The mean number of eggs laid per female mated to either control males (dashed line), iab-6^cocu males (solid line), or DTA-E males (grey dot dashed line) over a 10 day period. Mates of iab-6^cocu males lay normal numbers of eggs during the first day after mating (WRST p = 0.300) but lay significantly fewer eggs over 10 days when compared to mates of control males (rmANOVA p = <0.0001*, Control N = 51, iab-6^cocu N = 45, DTA-E N = 17). The drop in egg laying for controls on day 5 is atypical and was likely a response to food quality. B) The mean hatchability (#progeny/#eggs) per female for mates of control, iab-6^cocu, and DTA-E males for days 1–4 of the egg laying results reported in (A). Days 5–10 were omitted because iab-6^cocu mated females do not lay enough eggs on these days for analysis. Mates of iab-6^cocu males have comparable hatching totals for the eggs that they do lay when compared to mates of control males (WRST p = 0.37). Because DTA-E males do not produce sperm, their mates are expected to show zero hatchability. (Values greater than 1 represent instances where the number of progeny produced exceeded the number of eggs counted; this under-counting can result when females lay eggs under bubbles in the medium or directly on top of previously laid eggs. Hatchability values were not normalized to 1 so as to accurately report counter error.) C) The percentage of mated females willing to mate within 1 hour of exposure to a wild type male at 1, 4, and 10 days after an initial mating. Both groups of females initially mated to iab-6^cocu or control males are unreceptive (WRST p = 0.21, control n = 20, iab-6^cocu n = 26). At 4 d ASM females initially mated to iab-6^cocu males are significantly more receptive to courting males compared to mates of control males (WRST p = <0.0001*, control n = 21, iab-6^cocu n = 26). By 10 d ASM there is no difference between mates of control or iab-6^cocu males (WRST p = 0.84, control n = 19, iab-6^cocu n = 22).

A) For sperm competition assays cn bw females were first mated to either control (left) or iab-6^cocu (right) as the first male and allowed to mate a second time with a cn bw male. The proportion of progeny sired by iab-6^cocu males when acting as the first male (P1, # progeny from first male/total progeny) was significantly higher when compared to females who first mated with control males (WRST p = 0.038*, control N = 74, iab-6^cocu N = 98). B&C) Counts of sperm stored in mates of control (black) and iab-6^cocu (grey) males at 2 h, 4 d, and 10 d ASM. B) Mates of iab-6^cocu males have wild type numbers of sperm present in the seminal receptacle at 2 h (WRST p = 0.10, control N = 8, iab-6^cocu N = 11) and 4 d ASM (WRST p = 0.96, control N = 10, iab-6^cocu N = 8) but fewer at 10 d ASM (WRST p = 0.017*, control N = 19, iab-6^cocu N = 12) when compared to mates of control males. C) Mates of iab-6^cocu males show wild type numbers of sperm stored in the spermathecae at all time points. 2 h (WRST p = 0.13, control N = 7, iab-6^cocu N = 10); 4 d (WRST p = 0.38, control N = 10, iab-6^cocu N = 7); 10 d (WRST p = 0.77, control N = 17, iab-6^cocu N = 16).

Western blots using antibodies of known LTR-associated Acps CG9997, CG1656, CG1652, and CG17575 as well as STR Acp ovulin (Acp26Aa). Accessory gland extracts from a single control (lane 1) and iab-6^cocu male (lane 2) were used as positive controls and reproductive tract extracts from 4 virgin females (lane 9) were used as a negative control. Extracts from the reproductive tracts of females mated to control (+) or iab-6^cocu (Δ) were collected at 15′ (lanes 3–4, 2 RTs per), 30′ (lane 5–6 s, 3 RTs per), and 1 h ASM (lanes 7–8, 6 RTs per). A) Full length CG9997 is produced by iab-6^cocu males but is not present in the reproductive tracts of their mates. The smaller processed form of CG9997 is present in mates of iab-6^cocu suggesting that CG9997 is transferred. Both CG1656 and CG1652 are transferred to females normally by iab-6^cocu males, but both of these proteins run at a lower apparent molecular weight than in control males. B) iab-6^cocu males transfer more CG17575 to their mates than control males. Tubulin was used as a loading control for the female reproductive tracts. C) Both mates of control and iab-6^cocu males receive ovulin. However, the ovulin produced by iab-6^cocu males also runs at a lower apparent molecular weight than in controls.

PNGase F assays were used to examine glycosylation states of CG1656 and ovulin or determine whether N-linked glycosylation differences underlie the gel differences seen between control (+) and iab-6^cocu (Δ) males. Untreated (lanes 1–2) and PNGase F treated (lanes 3–4). In both cases the gel mobility differences seen between control and iab-6^cocu males are absent after PNGase F treatment suggesting that in iab-6^cocu males both proteins are improperly glycosylated.

Bold arrows delineate the new findings in this paper. The Drosophila male accessory gland consists of two secretory cell types: main cells and secondary cells. Main cells produce seminal proteins essential for inducing post-mating responses (PMR) in mated females; the function (if any) of secondary cells was unknown. Sex Peptide (SP) from the main cells induces many aspects of PMR, but persistence of its effects requires other seminal proteins (the LTR machinery: CG9997, CG17575, CG1652 and CG1656), whose cellular source was unknown. Here, we showed that the Hox gene Abd-B is essential for normal development of the secondary cells, but with regulatory characteristics (not shown in this figure) that differ from those used in segment identity. By deleting the secondary-cell regulatory element of Abd-B, we obtained males with abnormal secondary cells. Mates of those males failed to maintain the PMR, indicating that secondary cells play an essential role in reproduction: their products, along with main cell products, allow persistence of SP in mated females, thus prolonging their post-mating responses.