The insect brain is the center of developmental control and serves as the repository of the diapause program . In pupal diapause species, photoperiodic signal is perceived by larval brain during diapause induction. Then gene expression changes affected by photoperiod are first present in diapause preparation phase which follows diapause induction to regulate specific metabolism for diapause [3, 6, 22]. It is well known that after pupation, a shut-down of prothoracicotropic hormone (PTTH) in the brain and ecdysteroids in the prothoracic gland cause diapause initiation . Meola and Adkisson demonstrated that the shut-down of PTTH is found in day 0 of pupal brain of Helicoverpa zea, a closely related species to H. armigera . Thus, these differentially expressed genes isolated from the two libraries in day 1-2 pupal brain of H. armigera for diapause initiation are in response to hormones, but not photoperiodic signal.
In H. armigera, the photosensitive stage for diapsuse induction is from 5th instar to early stage of 6th instar. This is little different compared to H. armigera population from Okayama (Japan), whose photosensitive stage for diapause induction is the early fifth instar . After pupation, H. armigera diapause-type pupae are transferred into L14:10D photoperiod, all pupae will enter diapause, and all pupae will develop without diapause even if nondiapause-type pupae are transferred into L10:14D photoperiod (data not shown). Apparently, photoperiod regime does not affect pupal diapause or development.
The most remarkable characteristic of insect diapause is strong metabolic suppression. For example, in diapausing pupae of the flesh fly, Sarcophaga argyrostoma, the metabolic rate is approximately 90% lower than in nondiapause counterparts . Therefore, diapause was thought to represent a shutdown in gene expression. However, Joplin et al.  and Flannagan et al.  demonstrated that diapause should be a unique developmental pathway rather than a simple shutdown of gene expression. Recently, the proteomic analysis of the brain at diapause initiation has been reported, suggesting that the expression of many diapause-specific genes in the brain accompanies certain down-regulated genes [10, 11]. Thus, identification of diapause-associated genes at diapause initiation is the first step to understand the complex process of diapause. In the present paper, we isolated 304 diapause-specific mRNAs from H. armigera brain using SSH, and the subset of these genes with sequences similar to known genes in GenBank were classified according to their functions. Furthermore, we evaluated their mRNA expression at diapause initiation by RT-PCR and Northern blot analysis, and investigated the expression patterns of four important genes by RT-PCR and Western blot analysis, showing that these genes may be associated with diapause initiation.
From the SSH F library, we found a high percentage of undescribed sequences (61.3%). Some sequences may correspond to 3' or 5' untranslated regions (UTRs), so it is impossible to find their homologues in protein databases. However, most of these undescribed sequences can be classified as novel genes related to H. armigera pupal diapause initiation, because only a few genes related to developmental arrest have been identified. The large percentage of unknown (novel) genes in the F library shows that diapause is a complex physiological process involving a number of unknown genes in the regulation of developmental arrest.
We also constructed an R library to identify specific genes expressed in nondiapause individuals. The up-regulated gene expression in nondiapause pupae identified from the R library usually corresponded to down-regulated expression in diapause-type pupae (Figure 3A), so these genes from the R library will help us to identify the genes associated with insect diapause if these differentially expressed genes in diapause-destined pupae are further characterized. A total of 150 sequences from the two libraries that were homologous to known genes were obtained. According to gene ontology analysis, most genes belonged to cellular process and metabolic process in the category of "biological process"; this implies that the insect brain at diapause initiation focuses on alteration of cellular and metabolic state. Signaling and transcriptional regulator activity also showed significant differences between the two libraries. Up-regulation of signaling genes and down-regulation of transcriptional regulators at diapause initiation indicate that signaling pathways are changed, global transcription levels are down-regulated, and diapause does require a unique gene expression regulatory mechanism.
The quality and reliability of the two SSH libraries were validated by investigating gene expression difference between diapause- and nondiapause-destined pupae. The two libraries were quite reliable, so the SSH method was useful to search for genes related to pupal diapause initiation. Subsequently, the expression patterns of four genes were detected by RT-PCR and Western blot analysis. All four genes were expressed higher at both the mRNA and protein levels during early pupal development in diapause-destined individuals than their nondiapause-destined counterparts. Apparently, these genes from the SSH library may reflect differential expression between diapause- and nondiapause-destined pupae for promoting diapause initiation.
Based on the functions of the putatively up- and down-regulated genes (Table 1), we have proposed a possible mechanism for diapause initiation.