Disruption of the protein kinase N gene of drosophila melanogaster results in the recessive delorean allele (pkndln) with a negative impact on wing morphogenesis

Abstract

We describe the delorean mutation of the Drosophila melanogaster protein kinase N gene (pkn(dln)) with defects in wing morphology. Flies homozygous for the recessive pkn(dln) allele have a composite wing phenotype that exhibits changes in relative position and shape of the wing blade as well as loss of specific vein and bristle structures. The pkn(dln) allele is the result of a P-element insertion in the first intron of the pkn locus, and the delorean wing phenotype is contingent upon the interaction of insertion-bearing alleles in trans. The presence of the insertion results in production of a novel transcript that initiates from within the 3' end of the P-element. The delorean-specific transcript is predicted to produce a wild-type PKN protein. The delorean phenotype is not the result of a reduction in pkn expression, as it could not be recreated using a variety of wing-specific drivers of pkn-RNAi expression. Rather, it is the presence of the delorean-specific transcript that correlates with the mutant phenotype. We consider the delorean wing phenotype to be due to a pairing-dependent, recessive mutation that behaves as a dosage-sensitive, gain of function. Our analysis of genetic interactions with basket and nemo reflects an involvement of pkn and Jun-terminal kinase signaling in common processes during wing differentiation and places PKN as a potential effector of Rho1's involvement in the Jun-terminal kinase pathway. The delorean phenotype, with its associated defects in wing morphology, provides evidence of a role for PKN in adult morphogenetic processes.

Keywords: JNK pathway; PKN; Rho effector; signal transduction; wing morphology.

Figure 1

Characterization of the delorean wing mutation. (A) w¹¹¹⁸; pkn^dln/pkn⁺ heterozygotes have normal wings. Scale bar = 0.35 mm. (B, C) w¹¹¹⁸; pkn^dln/ pkn^dln homozygotes have wings that curve ventrally. The extent of curvature is variable. Bars = 0.35 mm. (D) Dissected wing of w¹¹¹⁸; pkn^dln/ pkn⁺ heterozygote. Bar = 0.2 mm. (E) Dissected wing of w¹¹¹⁸; pkn^dln/ pkn^dln homozygote shows creases from being mounted on a flat glass slide, longitudinal vein defects (arrow), and anterior margin defects. Bar = 0.2 mm. (F) Anterior wing margin of a w¹¹¹⁸; pkn^dln/pkn⁺ heterozygote has regularly spaced stout bristles (white arrow), slender bristles (black arrowhead), and recurved bristles (black arrows). Bar = 0.1 mm. (G) The anterior wing margin of a yw¹¹¹⁸; pkn^dln/ pkn^dln homozygote lacks many stout and recurved bristles and instead has gaps of bare marginal tissue. Bar = 0.1 mm.

Figure 2

Molecular map of the pkn gene of Drosophila melanogaster and relative insertion positions of pkn alleles examined. (A) A molecular map of the pkn gene was generated using annotated data from Flybase (http://flybase.org/reports/FBgn0020621.html). We used all available annotated transcripts to compile the intron-exon structure of the pkn gene. We used the evidence ranking of all potential transcripts to estimate the most likely molecular structure of pkn with more strongly supported exons depicted with darker shading. There are four potential transcription start sites each with a unique translation start. Predicted exon splicing patterns for these transcripts are shown below the molecular structure and potential alternative splicing is shown above. All products of pkn transcription would be predicted to have exons that code for the kinase domain (darkest shading). The pkn gene encompasses approximately 22 kb and is oriented such that transcription is toward the centromere (minus orientation). In order to adjust the gene structure to a reasonable length, some introns are not shown in their entirety (intron gaps are noted with // marks). The position (in the fifth intron), of the previously characterized P-element insertion allele pkn⁰⁶⁷³⁶ described as an amorphic allele (Lu and Settleman 1999), is shown relative to the insertion of the P[lacW] element in intron 1 that causes the delorean mutation. (B) Enlargement of the region of intron 1 that is impacted by the P-element insertion alleles characterized in this study. All of these insertions are within a ~140-bp region of intron 1. The pkn^k11209 allele is caused by a P[lacW] insertion at genomic position 5172332 (8bp duplicated target ATCTGAGC); the pkn^dln allele is caused by a P[lacW] insertion at genomic position 5172249 (8-bp duplicated target GTTTAACC); the pkn^rG232 allele is caused by an P[PZ] insertion at genomic position 5172186 (8-bp duplicated target GGCCGTGC).

Figure 3

Expression of the mini-white⁺ reporter of pkn^dln is influenced by the pkn insertion allele in trans. (A) Eye phenotypes of 7- to 10-day-old flies are shown. Males are shown on the left, females are shown on the right. Flies on top are genotype pkn^dln/pkn^dlnΔ5 whereas flies on the bottom are pkn^dln/pkn⁺. Despite the sex-specific differences in expression of the mini-white⁺ reporter, it can be seen that even though all flies contain only one dose of the mini-white⁺ reporter, flies with the genotype pkn^dln/pkn^dlnΔ5 have increased pigmentation. (B) This same increase in pigmentation can be seen even when the pkn insertion allele in trans lacks sequence in common with the mini-white⁺ reporter of pkn^dln. Five-day-old males with the genotype pkn^dln/pkn^rG232 (top) or genotype pkn^dln/pkn⁺ (bottom) are shown.

Figure 4

The pkn^dln allele generates a novel transcript. (A) Rapid amplification of 5′ cDNA ends (5′ RACE) products derived from pupal RNA. Left to right: pkn^dln/pkn^dln homozygous products; pkn^dln/Df(2R)w45-30n hemizygous products; pkn⁺/Df(2R) hemizygous product; 100-bp molecular weight standard (Promega). (B) Sequence of the 5′ end of the pkn transcriptional unit. Only one strand of DNA is shown. Exonic bases are uppercase; nonexonic bases are lowercase. The putative TATA box is boxed as is the predicted translational start site. The 5′-most nucleotide of various known pkn cDNAs are bolded. The positions of the three intron 1 P-element insertions described in this article are indicated with arrows (top to bottom: pkn^k11209, pkn^dln, pkn^rG232). Intronic bases that are included in the pkn^dln-specific transcript are orange. (C) Sequence of the 5′ end of the pkn^dln-specific transcript. Exonic bases are uppercase; intronic bases are lowercase. The portion derived from the 3′ end of the P[lacW] transposon is highlighted in green. Bases that are excised from pkn⁺ transcripts but are included in the pkn^dln transcript (corresponding to the orange sequence shown in part 4B) are highlighted in orange. The predicted translational start site is boxed.

Figure 5

The delorean mutation interacts with basket but not with LIMK1. (A, A′) Dissected wing of pkn^dln/ pkn^dln shows the delorean phenotype. (B, B′) Dissected wing of bsk¹ pkn^dln/ pkn^dln shows a more severe delorean phenotype. (C, C′) Dissected wing of LIMK1^EY08757/LIMK1⁺; pkn^dln / pkn^dln shows the delorean phenotype. See Table 3 for quantitative analysis. Left panels (A, B, C) are whole wings, scale bar = 0.2 mm. Right panels (A′, B′, C′) are close-ups of anterior wing margins, bar = 0.1 mm.

Figure 6

The delorean mutation interacts with nemo. (A, A′) Dissected wing of pkn^dln/ pkn^dln ; nmo^P/+ shows a more extreme delorean phenotype than wings of pkn^dln siblings without nmo^P. Gaps in crossveins are indicated by arrows. (B, B′) Dissected wing of pkn^dln/ pkn^dln shows the delorean phenotype. (C, C′) Dissected wing of pkn^dln/pkn⁺ ; nmo^P/nmo⁺ shows a wild-type phenotype. Left panels (A, B, C) are whole wings, scale bar = 0.2 mm, and right panels (A′, B′, C′) are close-ups of anterior wing margins, bar = 0.1 mm.

Figure 7

Pairing dependence of pkn^dln expression as a working model to explain the delorean phenotype. The model depicts the 5′ end of the pkn gene showing only the affected region of the gene (first two exons and intron 1; refer to Figure 2). Thickness of arrows represents assumed relative levels of transcription. Transcription of the wild-type transcript (pkn⁺) begins at the first exon. Initiation of the delorean-specific transcript, referred to as pkn^dln, from within the P-element generates a transcript that is either expressed ectopically or regulated differently than pkn⁺. Relative levels of pkn⁺ and pkn^dln transcripts are given with the “+” symbol. The schematic in the far right column depicts the position and shape of the wing for each genotype. (A) A reduction in the expression of wild-type pkn and an increase in the expression of pkn^dln, occurs when the pkn^dln allele is homozygous. The level of pkn^dln expression is elevated as a result of pairing thereby causing the composite wing defects of the delorean phenotype. The level of wild-type transcript may be reduced due to the additional sequence in the form of the P-element in the first intron that must be removed by splicing. (B) Levels of the delorean-specific transcript are reduced in pkn^dln/pkn⁺ heterozygotes due to the absence of P[lacW] in trans to the pkn^dln allele. The reduction in the level of pkn^dln transcript is below a threshold needed to generate a phenotype. (C ,D) Our analysis of heteroallelic combinations of pkn alleles also lead us to the conclusion that a reduction in the level of delorean-specific transcript occurs in the absence of either any homology-based pairing in trans (C; pkn^dln/Df(2R)) or when pairing is disrupted (D; pkn^dln/pkn^dln; Dp(2:3)). In the case of pkn^dln/Df(2R) there would be a reduction in the level of both pkn^dln and pkn⁺ transcripts. Reduced levels of pkn^dln transcript can suppress the delorean phenotype. In the case of pkn^dln/pkn^dln; Dp(2:3), the transposed sequence present in the duplication (Dp) is able to disrupt pairing between the pkn^dln alleles due to homology resulting in decreased pkn^dln transcript levels. It is also likely that the level of wild-type pkn transcription is affected by the pairing state at the pkn gene. We consider it reasonable to infer that the relative levels of a delorean-specific transcript are involved in determining the extent of the delorean phenotype.