Drosophila S6 Kinase: A Regulator of Cell Size

See allHide authors and affiliations

Science  24 Sep 1999:
Vol. 285, Issue 5436, pp. 2126-2129
DOI: 10.1126/science.285.5436.2126


Cell proliferation requires cell growth; that is, cells only divide after they reach a critical size. However, the mechanisms by which cells grow and maintain their appropriate size have remained elusive.Drosophila deficient in the S6 kinase gene (dS6K) exhibited an extreme delay in development and a severe reduction in body size. These flies had smaller cells rather than fewer cells. The effect was cell-autonomous, displayed throughout larval development, and distinct from that of ribosomal protein mutants (Minutes). Thus, the dS6K gene product regulates cell size in a cell-autonomous manner without impinging on cell number.

In unicellular and multicellular organisms, specific cell types have a characteristic size (1). In multicellular organisms, the control of cell size has the added importance of dictating the proportions of organs, limbs, and the living entity (1). Therefore, mechanisms must exist to integrate cell growth and proliferation, so that cells maintain an appropriate size (1). In yeast, a minimum cell size must be achieved before cells initiate S phase and divide, whereas mutations that block cell cycle progression do not inhibit cell growth (2). Studies in which cell cycle regulators were manipulated in the imaginal wing disc of Drosophila melanogaster(3) largely confirmed these observations in metazoans. However, little is known concerning the identity of regulatory components and signaling pathways that control cell growth or the mechanisms by which they are integrated with the control of cell proliferation in the developing organism (1,4). The mammalian 40S ribosomal protein S6 kinases [p70 S6 kinase, also called S6K1, and the very similar S6K2 (5)] may represent such regulatory components (3). These enzymes serve as the physiological kinases for the ribosomal protein S6 (5). Their role in cell growth stems from their function in controlling the increased translation of mRNAs encoding for ribosomal proteins (6). InDrosophila, a family of mutants defective in ribosomal protein production, termed Minutes, display short and slender bristles and a delayed developmental program, which results from a slower rate of cell growth and division (7). Because S6K1 regulates ribosomal protein production in mammals, loss ofDrosophila S6K (dS6K) function could have a direct impact on cell growth and proliferation.

We found that a female sterile mutant, fs(3)07084(8), contained a P-element insertion in the 5′ noncoding region of the dS6K gene (Fig. 1A) (9). Only 25% of the expected number of homozygous fs(3)07084(dS6K07084 ) flies emerged as adults, with a 3-day delay and reduced body size. This phenotype was rescued either by excision of the P element or by a dS6K or mammalian S6K transgene (10). Northern (RNA) blot analysis (Fig. 1B) (11) and sequencing of a reverse transcriptase polymerase chain reaction product (RT-PCR) from homozygous mutant flies revealed the presence of anomalous transcripts (Fig. 1C) (11), suggesting that dS6K expression may persist in homozygous dS6K07084 flies. More severe alleles were generated by imprecise P-element excisions (12), removing part of the dS6K gene. Most of these flies died as larvae, with the lethality rescued by expression ofdS6K or mammalian S6K transgenes (10). One of the excisions, dS6Kl-1 , removed part of the first exon, including a portion of the catalytic domain (Fig. 1A) (12). The few surviving dS6Kl-1 homozygous flies emerged after a 5-day delay, lived no longer than 2 weeks, and displayed a severe reduction in body size, with all body parts apparently affected to the same extent (Fig. 1D). Thus, loss of dS6K function induces female sterility, a strong developmental delay, a severe reduction in growth, and often death.

Figure 1

Identification of P-element–induced mutation in the dS6K gene. (A) dS6Kgene showing introns and exons. The dS6K cDNA sequence has recently been modified (GenBank accession number U66562). The P-element (PZ) insertion point and the coding sequence (black boxes) are indicated. The mRNA structure of thedS6Kl-1 deficiency is represented at the bottom. Single- letter amino acid abbreviations are as follows: A, Ala; F, Phe; K, Lys; L, Leu; M, Met; V, Val; and Y, Tyr. (B) Northern (RNA) blot analysis with the dS6K cDNA as a probe (11): (lane 1) three transcripts of 3.3, 4.4, and 5.8 kb detected in total RNA (20 μg) isolated from ovaries of homozygousdS6K07084 females or (lane 2) three transcripts of 2.8, 3.7, and 5.0 kb in the same amount of total RNA isolated from wild-type females. (C) Sequence of RT-PCR fragment generated from mRNA of homozygousdS6K07084 female ovaries with two oligonucleotides covering part of the P element and the second exon of dS6K (11). The sequences derived from the P element and the second exon of dS6K are in bold. Oligonucleotides used for PCR reaction are underlined, and the ATG translational start codon for dS6K is boxed. (D) Comparison of wild-type (+/+) anddS6Kl-1 (l-1/l-1) homozygous females 4 days after emergence. Body weights of wild-type anddS6Kl-1 homozygous females were 1.49 ± 0.04 mg and 0.80 ± 0.1 mg, respectively.

As mammalian S6Ks control the synthesis of ribosomal proteins (6), we hypothesized that thedS6Kl-1 phenotype might be equivalent to that of Minutes (7). The Minute M(3)95A, harbors a P-element insertion that severely reduces the expression of ribosomal protein S3 (7). However, analysis of M(3)95A (Fig. 2A) and two other Minutes showed no effect on size (10), although all displayed a developmental delay and slender bristles. In contrast, the bristles of homozygousdS6Kl-1 flies were proportional to body size (Fig. 2B). To determine whether the reduction in body size of homozygous dS6Kl-1 flies was due to a decrease in cell number, we compared cells in wings and ommatidia in eyes of wild-type and dS6K mutant flies. The cell density was greater in wings of homozygousdS6Kl-1 flies (as represented by each hair) than in wild-type flies (Fig. 2C) (13). The difference in cell size was almost 30%, and flies homozygous for partial loss of function dS6K07084 displayed an intermediate cell size (Fig. 2C). However, the total number of cells in wings remained constant (Fig. 2C). Analysis of eyes revealed a similar phenotype with reduced size but no effect on the number of ommatidia (Fig. 2D) (13). Thus, in dS6K mutants, the decrease in the rate of proliferation (see below) is probably attributable to a reduction in ribosomal protein synthesis, whereas the effect on cell size may be due to the absence of S6 phosphorylation and an altered pattern of translation (6).

Figure 2

Loss of dS6K function affects cell size. (A) Comparison of wild-type (+/+) and heterozygousM(3)95A (M95A) females 4 days after emergence. (B) Comparison of dorsal bristles at the scutellar-notum junction of wild-type (+/+), heterozygousM(3)95A (M95A), and homozygous dS6Kl-1 (l-1/l-1) flies. (C) (top) Wings (13) from wild-type and homozygous dS6Kl-1 flies. (middle) Higher magnification of top panels (13). (bottom) Cell size and total cell number of wild-type (+/+), homozygousdS6K07084 (P/P), and homozygousdS6Kl-1 (l-1/l-1) flies. The bottom right panel represents the calculated number of intervein cells in the dorsal wing blade (13). Values are means of eight different wings. (D) Scanning electron micrographs of eyes from wild-type (+/+) and homozygousdS6Kl-1 (l-1/l-1) flies (13).

The reduction in cell size of dS6Kl-1 flies indicates either that cells are proliferating at a smaller size or that flies emerge from the extensive developmental delay before completion of the last round of cell growth. To examine these possibilities, we analyzed proliferating epithelial cells from the imaginal wing disc of larvae at the end of the third instar (14). Imaginal discs give rise to the adult structures (14). At the end of the third instar, wing disc cells still require two mitotic cell cycles before they differentiate (15). Comparison of wing discs from homozygousdS6Kl-1 and wild-type larvae revealed that mutant discs were substantially smaller in size (Fig. 3A) (14). Analysis of single cells from discs with a fluorescence-activated cell sorter (FACS) confirmed that, on average, cells derived from dS6Kmutants were smaller than wild-type cells (Fig. 3B). There was no apparent difference between the distributions of dS6K mutant and wild-type cells within each phase of the cell cycle (Fig. 3C), implying that the dS6Kl-1 loss-of-function mutation affected all stages of the cell cycle. Analysis of disc cells during puparium formation, when proportionally more cells are present in G2 phase (16), also showed no detectable difference in the cell cycle distribution of mutant and wild-type cells (Fig. 3D). In addition, the number of wing disc cells present in somatically induced clones, marked by ectopic expression of β-galactosidase, was reduced in mutant versus wild-type larvae. Consistent with this, cell cycle times were 12.5 ± 1 hours and 24 ± 4 hours for wild-type and mutant wing disc cells, respectively (Fig. 3E). Thus, loss of dS6K function leads to cell proliferation at a smaller size and at a reduced rate, without affecting any specific stage of the cell cycle.

Figure 3

Cell cycle progression at reduced rate and size in loss-of-function dS6K mutants. (A) Comparison of wild-type (+/+) and dS6Kl-1 (l-1/l-1) homozygous imaginal wing discs of developing third instar larvae at the end of the wandering stage. Posterior, right; dorsal, up. (B) Imaginal wing disc cell size, as measured by forward scatter (FSC). (C) DNA content of imaginal wing disc cells measured by FACS analysis. (D) Same as (C), except that wing disc cells were analyzed from early prepupal stage. (E) Clones in wing disc of wild-type (+/+) anddS6Kl-1 (l-1/l-1) larvae were detected by β-galactosidase antibody staining 53 hours after induction of somatic recombination (3). The number of wing disc cells present in 80 clones each of wild-type and mutant larvae was counted, and cell cycle times were determined (3).

S6Ks have been implicated in the synthesis of mitogens (17); therefore, dS6K mutants may affect cell size through the loss of a humoral factor that regulates cell growth. To examine this possibility, we generated genetically marked homozygous mutant cells in a heterozygous mutant background by somatic recombination (18). At the wing margin, homozygousdS6K mutant sensory bristles, identified by ayellow (y ) marker (Fig. 4A) (13, 18), were reduced in size compared with their neighbors. In eyes, homozygousdS6K mutant photoreceptor and pigment cells were marked by awhite (w ) mutation and recognized by the absence of red pigment, appearing as dark spots in photoreceptor cells (Fig. 4B). Again, only mutant cells were reduced in size, indicating that dS6K acts in a cell-autonomous manner. Because dS6K mutations affect size in a cell-autonomous manner, expression of an extra copy of the wild-type gene in a specific compartment might positively affect growth. A compartment represents an independent unit of growth and size control, thought to be analogous to a mammalian organ (1). The wing disc is composed of a dorsal compartment and a ventral compartment that fold in an apposed manner at the wing margin to generate the flattened wing blade (19). Because the apterous promoter is only functional in the dorsal compartment of the wing disc (20), it was coupled to the GAL4 transcription factor to induce an extra copy of thedS6K gene linked to a UAS responsive element (21). An increase in cell size of less than 1% should alter the morphology of the adult wing blade (22). In all UASdS6K lines we examined, dS6K protein expression was increased and the wing was convex and bent downward (Fig. 4C). The phenotype can be explained by an increase in the size of the dorsal versus the ventral wing surface, forcing the wing blade to curve down to accommodate the greater surface (22). Therefore, increased expression of dS6K positively affects growth in a cell-autonomous and compartment-dependent manner.

Figure 4

Cell and compartment autonomous actions of dS6K. (A) Homozygous mutant sensory bristles at the wing margin and (B) homozygous mutant cells in the eye were induced by somatic recombination (18). (A) A ymarked dS6K mutant sensory bristle is indicated by the arrow. (B) The w+ wild-type and w marked homozygous dS6K mutant ommatidia are recognized by the presence (large box) or absence (small box) of red pigment, respectively. (C) Comparison of GAL4-apterousflies in the absence or presence of a transgene harboring a UAS response element coupled to an extra copy of dS6K. The insets show antibody staining of dS6K in imaginal wing discs, and the arrow indicates the dorsal-ventral compartment boundary (20, 21). Posterior, right; dorsal, up.

S6Ks appear to be a downstream effectors of the phosphatidylinositide-3OH kinase (PI3K) signaling pathway (23). However, activated or dominant interfering alleles of PI3K affect cell number and cell size (24). This would imply that S6Ks reside on a branch of the PI3K signaling pathway that controls cell growth and size but not cell number. Overexpression of the cell cycle regulator E2F in the posterior compartment of the wing disc increases cell number without affecting final compartment size (3). These findings are consistent with the hypothesis that compartments, like organs, adjust their final mass independent of cell number (1). However, the dS6K phenotypes described here suggest that cell size participates in the control of compartment size. Indeed, S6Ks are thought to play a critical role in organ hypertrophy, where the organ increases in size as a function of demand (25).

  • * These authors contributed equally to this work.

  • Present address: Department of Zoology, North Dakota State University, Fargo, ND 58105, USA.

  • To whom correspondence should be addressed. E-mail: gthomas{at}


Stay Connected to Science

Navigate This Article