BIOLOGICAL OVERVIEW

Pigment-dispersing hormones (PDH) are neuropeptides secreted from the sinus gland of the eyestalks of crustaceans. Their physiological functions in crustaceans include translocation of retinal distal pigments and epithelial chromatophoral pigment dispersion (Rao, 1993). All PDHs identified so far are octadecapeptides, and their C-terminals are invariably modified by alpha-amidation. Because their physiological roles have not been clearly defined in insects, they are known as pigment-dispersing factors (PDFs). The insect peptides have substantial pigment-dispersing activies when assayed on the eyestalkless fiddler crab. However, PDFs are unlikely to play a role in cuticular pigment migration, since insects do not possess the epithelium chromatophore found in crustaceans (Park, 1998 and references therein).

Cloning the Drosophila Pdf gene is a first step toward unraveling its biological function. The Pdf gene has been cloned using RACE (rapid amplification of cDNA ends). The design of the degenerate primer used for cloning was based on the amino acid sequence homology among members of the PDH family. Adult heads were chosen as the source of RNA, since immunocytochemical studies have shown more PDH-immunoreactive cells present in the adult Drosophila brain than in the abdominal ganglion (Park, 1998).

Several lines of evidence suggest that PDFs might play a role in the regulation of insect biological rhythms. Immunocytochemical studies, using antibodies against crustacean PDHs, reveal three groups of PDH-immunoreactive neurons in the brain of orthopteroid insects. One group of cells, known as PDFMe for its location at the anterior edge of the medulla, fulfills the anatomical criteria proposed for the cardiac pacemakers in these insects. Lesion studies, along with immunocytochemistry and behavioral analysis in the cockroach, suggest that PDFMe might be part of the circadian pacemaker in this insect (Stengl, 1994). Cell bodies of Drosophila PDH-immunoreactive neurons are located at the anterior base of the medulla; their axonal arborizations are well suited to modulate the neuronal activities in the optic lobe and relay the circadian information to the midbrain (Helfrich-Foster, 1995 and 1997). In adults, this group consists of four to six neurons with large somata (large PDFMe neurons) and four neurons with small somata (small PDFMe neurons). Both the small and the large PDFMe neurons are identical to the ventral lateral neurons, a group of neurons containing the Period protein (Helfrich-Forster, 1997). beta-PDH-like antigens colocalize with those detected by a new antibody to distinct sequences of the Drosophila proPDF precursor (anti-PAP). Pdf gene products are found in three cell types. These include the LN_vs, two to four tritocerebral cells (PDF-Tri), and four to six abdominal cells (PDF-Ab) (Renn 1999).

An evaluation was made of the activity rhythms of flies with PDF cell ablations. Ablation was accomplished by ectopic expression of the cell death genes reaper or head involution defective in cells that normally express Pdf. No PDF-positive cell bodies are detected in the CNS of third instar larvae bearing pdf-GAL4 and UAS-rpr. In some of these larval CNSs, however, a few residual processes in the dorsal brain are very faintly stained. In adults bearing both transgenes, neither s-LN_v cell bodies nor dorsal processes are stained, whereas some l-LN_vs are stained. p35 encodes a caspase inhibitor that can rescue rpr- or hid-mediated cell death. When p35 is coexpressed with rpr, most of the larval LNs (80%), adult s-LN_vs (70%), and l-LN_vs (80%) survive. The PDF-Tri neurons (which normally cease to express pdf after adult days 1-2) remain persistently PDF immunoreactive in rpr-rescued brains, up to adult day 10. This suggests that PDF-Tri cells normally die in young adults (Renn, 1999).

Analysis of DD behavior shows that 63% of rpr-ablated flies are rhythmic for the entire period, whereas only 17% of hid-ablated flies sustain such rhythmicity. As in the case of pdf⁰¹, separate periodogram analysis of DD days 3-9 reveals decreased proportions of rhythmic individuals in cell-ablated flies. The subnormal SNR values computed for both the rpr- and hid-ablated flies are consistent with their abnormal free-running behavior. Finally, the rpr and hid ablation individuals that are persistently rhythmic in DD tend to manifest short circadian periods (Renn, 1999).

Despite having a slightly lower number of PDF cells, animals that coexpressed rpr and p35 display essentially normal behavior. The free-running period of the rescued group is approximately 0.5 hr longer than control values. In contrast to animals that coexpress p35 and rpr, only about 70% animals that coexpress p35 and hid are rhythmic in DD, and the SNRs for this group are intermediate between those for controls and for hid ablation. This incomplete behavioral rescue parallel the histological findings (Renn, 1999).

What then is the biological function, if any, of Pdf in the photoperiod response? Isolation of mutations in the Pdf gene has allowed this question to be addressed. The most severe phenotype displayed by pdf⁰¹ mutants and by PDF cell-ablated animals is that the majority are arrhythmic in constant darkness (DD). Both sets of animals are rhythmic over the first 1-2 days of constant darkness. Their locomotor patterns become arrhythmic gradually over a 9 day period. It is concluded that circadian behavior is largely independent of Pdf hormone and of LN_v neurons during normal light dark (LD) cycles and short term continuous darkness, but the requirements for Pdf and the cells that produce the hormone are revealed during artificially sustained constant DD conditions (Renn, 1999).

What features of LN_v neurons and pdf signaling could explain this phenotype? In cockroaches, injection of beta-PDH into the brain causes phase delays in daily locomotor activity, consistent with a role for the peptide in a nonphotic clock input (Petri, 1997). The morphology of l-LN_v neurons suggests a basis for how secreted Pdf gene products could access the pacemaker neurons (Kaneko, 1997). A subset of l-LN_v cells projects axons across the midline to the area containing the contralateral LN_v cell bodies. Therefore, rhythmic l-LN_v release of PDF could produce a phase delay in pacemakers of the opposite side and, thus, contribute to bilateral synchrony (Renn, 1999).

This scenario predicts that in Drosophila mutant for Pdf, the circadian clock will operate with advanced phase in LD and run more quickly in DD. These behaviors are the same as those observed for Pdf-null animals. The deterioration of free-running rhythmicity over DD days 1-3 may therefore reflect a gradual loss of synchronization between bilateral pacemaker centers. The disconnected mutant displays a progressive damping of rhythmicity and also lacks LN_v neurons (Helfrich-Forster, 1998). Likewise, the ablation of the avian pineal gland produces an analogous behavior in sparrows: when transferred from LD cycling to DD conditions, operated animals display a progressive loss of behavioral rhythmicity (Menaker, 1976). This effect has been shown to derive from lack of melatonin, which normally helps to maintain a mutual synchronization between the pineal and other pacemaker structures (Cassone, 1984).

The dorsally projecting s-LN_v cells may have a greater role in regulating circadian locomotor rhythms than the l-LN_vs. Cell ablation studies are consistent with the proposition that a single LN_v is competent to organize behavioral rhythmicity. This same conclusion was reached in an analysis of disco mutants (Helfrich-Forster, 1998). Likewise, the circadian regulators Clock and cycle regulate pdf expression in s-LN_vs, but not in l-LN_vs (J. H. Park. and C. Helfrich-Förster, et al., unpublished data cited in Renn, 1999). That result argues that neuropeptide expression by s-LNs neurons is especially important for the circadian behavioral regulation that has been inferred from analysis of pdf⁰¹ animals (Renn, 1999).

While the current results strongly support the hypothesis that LN_vs are critical circadian pacemaker neurons, analyses of both pdf⁰¹ and PDF-ablated flies reveal minorities of animals that maintain weak rhythmicity in DD. These low proportions of rhythmic individuals suggest the involvement of secondary pacemaker neurons involved in the circadian regulation of behavior. Their cellular identities are unknown, but on the assumption that they will express the clock genes period and timeless, three specific candidates are proposed. The first is a fifth per-positive, pdf-negative LN_v neuron; a cell with such properties was found in larvae, and it may also exist in adults; if so, its activities are presumed not to be affected in Pdf mutants. The second candidate cell type is represented by the dorsolateral LN_d cluster of neurons. The third plausible candidate cell type is represented by the Dorsal Neurons (DNs) of posterior-medial brain regions. The second and third candidate cell types are both Pdf negative (Helfrich-Forster, 1995). Free-running rhythms are more severely disrupted in disco flies than in pdf⁰¹ flies, and it is proposed that this is because disco flies lack almost all per-positive LN neurons, not just the ventral LN group (Renn, 1999 and references therein).

The pdf⁰¹ and neuron-ablated animals entrain to a 24 hr light:dark cycle and show considerable rhythmicity. This feature is noticeably different from other clock mutants, which are solely driven by photoperiod. This difference suggests the clock is still running in pdf⁰¹ and cell-ablated flies and that PDF is therefore not a central component of the clock mechanism. However, both pdf⁰¹ and cell-ablated flies display phase-advanced evening activity peaks in LD, and if rhythmic in DD, they display a short free-running period. It is proposed that the same physiological mechanism underlies both of these phenotypes. Both features are also displayed by per^Clk and norpA mutants. Light resets these fast-paced clocks by about 1 hr per day to 24 hr. Hence, pdf⁰¹, like per^Clk and norpA mutants, produces fast-paced clock movements in LD and DD. However, pdf⁰¹ variants have additional phenotypes: they fail to anticipate a lights-on transition in LD and are largely arrhythmic in DD. This suggests that a fast clock is not the only or appropriate explanation for all phenotypes associated with the pdf⁰¹-mutated and cell-ablated flies (Renn, 1999 and references therein).

The predicted Pdf gene product is a neuropeptide precursor, pro-PDF, which is presumed to be processed to two or more final peptide products that include the PAP and amidated 18-amino acid PDF molecules (Nassel, 1993). The pharmacological activities of injected beta-PDH peptide in other insects (Pyza, 1996; Petri, 1997) are consistent with the hypothesis that it represents a secreted agent. While the results of this study suggest that PDF is the principal circadian messenger in Drosophila, certain details remain ambiguous and will require further study. Two sets of results warrant comment. (1) A role for PDF neurons does did not implicate LN_v neurons exclusively. The lack of transmitter in PDF-Tri and PDF-Ab neurons, or their genetic ablation, may have contributed to the phenotypic defects. This is considered unlikely, as neither cell type expresses clock genes. Furthermore, PDF-Tri cells normally undergo apoptosis before the stage when locomotion is measured. (2) The experiments described here do not tell when the lesions studied have their effects: lack of transmitter or lack of Pdf neurons at an early, preadult stage may covertly affect behavioral periodicity, as well as having later physiological effects. The normal morphology of LN_v neurons in mutant animals argues against this possibility, and PDF neuropeptides have not previously been implicated in developmental functions. However, future experiments employing conditional manipulations will be necessary to evaluate this possibility (Renn, 1999 and references therein).

The role of Pdf in the Drosophila circadian system is notable, because it is a neuropeptide essential for circadian rhythm output. In rodents, the vasopressin neuropeptide gene is rhythmically expressed in SCN under the influence of the Clock gene (Jin, 1999). However, a functional role for vasopressin in SCN regulation of locomotor behavior has not been defined. Several neuropeptides can reset the phase of daily rhythms, but their effects appear to mimic natural inputs to circadian cycling, not its output. Norepinephrine is released from sympathetic nerve terminals in a circadian manner, but these neurons represent a distant, polysynaptic target of the neuronal output emanating from SCN. To date, PDF-related peptides have been found only in arthropods and mollusks. Whether related PDF-like peptides have analogous circadian functions in vertebrates, or whether a nonrelated transmitter has such functions, remains to be determined (Renn, 1999 and references therein).

GENE STRUCTURE

cDNA clone length - 1080

Bases in 5' UTR - 279

Exons - 1

Bases in 3' UTR - 492

PROTEIN STRUCTURE

Amino Acids - 102

Structural Domains

The putative 102-amino-acid precursor (prepro-PDF) consists of a signal peptide and a PDF-associated peptide, followed by the mature PDF. The PDF-associated peptide region of the precursor is highly diverged from those of the crustacean precursors, whereas the primary structure of the mature PDF is conserved in other members of the pigment-dispersing hormone family. The first 16 amino acids show features typical of signal peptides, responsible for directing the secretion of the protein. The next 63 amino acids (residues 17-79) are termed PDF-associated peptides (PAP). These are followed by the mature 18-amino-acid PDF (residues 83-100). A potential proteolytic processing site (RKR) is present between PAP and PDF at residues 80-82. As is observed for other members of the PDH family, the presence of a consensus alpha-amidation signal (Gly-Lys) suggests that PDF in Drosophila exists as an alpha-amidated peptide. Within the PAP sequence, there is another potential clevage site; dibasic amino acids at positions 35-36 (RK); usage of this site along with the tribasic one would yield peptides of 18 and 43 amino acids. From BLAST search data, the primary structure of the PAP region of the prepro-PDF has no sequence homology with any known polypeptide, except for the region of N-terminal 8 amino acids (residues 15-22): CCQWGYCG. This region is highly homologous (7/8) to parts of the putative carbohydrate recognition domains of pokeweed mitogen and other carbohydrate-binding proteins such as lectins and agglutinins. The significance of this similarity is unknown. In any case, none of the squences within Drosophila's PAP is similar to corresponding regions of PDH precursors identified in crustaceans. Despite this divergence of PAP, the deduced primary structure of the mature PDF in Drosophila is homologous to other PDH family members. The fly peptide differs from that of the cockroach by only three amino acids. There are two amino acid residues unique to the fly peptide. Since these amino acids are more hydrophilic than the corresponding ones in other PDHs, the fly's PDF is anticipated to be more hydrophilic than other PDH peptides (Park, 1998).

date revised: 12 March 2000

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.