The PIK3CA gene spans a total genomic size of 86,190 bases and is composed of 21 exons, 20 of them coding exons of varying lengths. Putative pseudogenes of PIK3CA have been described on chromosomes 16 (gi 28913054) and 22q11.2 (gi 5931525), the later one in the Cat Eye Syndrome region. These regions are highly homolog to the sequences of exons 9 and 11-13 of the PIK3CA gene.

The human PIK3CA transcript has an open reading frame of 3,207-bp, predicting a protein of 1,068 amino acid residues.

A pseudogene has been identified in chromosome 20 (LOC100422375) spanning exons 9 to 13 with 95% homology in both exons and introns. A non-specific amplification of PIK3CA could result in a E545A false positive mutation which is the amino-acid most commonly mutated in PIK3CA.

The PIK3CA gene encodes the p110alpha protein which is a catalytic subunit of the class I PI 3-kinases (PI3K). Class I PI3K are heterodimeric molecules composed of a catalytic subunit, a p110, and a regulatory subunit. There are three possible calatytic subunits p110alpha, beta or delta.

Widely expressed.

The p110alpha localizes in the cytoplasm.

Class I PI 3-kinases (PI3K) are one of the most prevalently dysregulated pathways in human cancer and it is linked to many cellular functions, including cell growth, proliferation, differentiation, motility, survival and intracellular trafficking. PI3K convert PI(4,5)P2 to PI(3,4,5)P3 on the inner leaflet of the plasma membrane. The PI(3,4,)P3 provokes the recruitment to cellular membranes of a variety of signalling proteins, containing PX domain, pleckstrin homology domains (PH domains), FYVE domains and other phosphoinositide-binding domains. One of these is the protein kinase B (PKB/AKT) a very well known protein that is activated as a result of its translocation to the cell membrane where it is then phosphorylated and activated by another kinase, called phosphoinositide dependent kinase 1 (PDK1). The phosphorylation of AKT leads to the activation of the TSC/mTOR pathway. Accordingly, PIK3CA and PTEN genetic alterations are usually alternative events.

Somatic mutations at the PIK3CA gene have been found in tumors and thus, it is a bona fide oncogene (Samuels et al., 2004). Sequencing analyses of pancancer genomes suggests that PIK3CA has at least 19 mutational hotspots (Chang et al, 2016). Most mutations occur in two of the five domains of p110alpha, the helical and kinase domains. However, while there are no frequent hotspot mutations within the Ras-binding domain (RBD), hotspot mutations within the adaptor-binding domain (ABD) and the C2 domain are also frequently found. Hotspots Arg38 and Arg88 are located within the ABD at the interface with the kinase domain. These mutations disrupt hydrogen bonds between the domains which may alter the conformation of the kinase domain and therefore its enzymatic activity. Similarly, hotspots Asn345 and Glu453 within the C2 domain occur at the interface with the iSH2 domain of p85 and are expected to alter the interaction between these domains. Within the helical domain, the most common hotspots Glu542 and Glu545 are usually mutated into lysine and are located at the interface with the nSH2 domain of p85. A structural model of the p110/niSH2 complex suggests that the contacts between these amino acids and the nSH2 domain occur within a region of the nSH2 domain which is also in contact with the kinase domain of p110. These contacts may serve as the mechanism by which the helical domain mutations alter the activity of the enzyme. In the kinase domain, His1047 is frequently mutated to arginine within a helix at the end of the activation loop. Another less frequent kinase mutation of Met1043 occurs within the same helix and likely has similar effects on enzymatic activity. Thr1025 mutations are located near the N-terminus of the catalytic loop which may directly alter the conformation of the catalytic loop as the mechanism by which the mutation alters enzymatic activity (Huang, 2008). Two other hotspot positions, Lys111 and Gly118, are located outside of the five domains in the 81 residue linker between ABD and RBD (Chang, 2016).
The distribution of PIK3CA hotspot mutations can be significantly variable across tumour types in the kinase and helical domain. For example, cervical and bladder cancers share similar distributions of hotspots but they have far fewer H1047R kinase mutations than breast cancer. Similarly, endometrial and colorectal cancers share distribution patterns but display more R88Q mutations than other tumor types, maybe because this change is a target mutation of POLE genomic instability. However, colorectal cancers differ from endometrial cancers in their clonality. In endometrial tumors, both E545 helical domain mutations and H1047 kinase domain mutations are early clonal mutations whereas in colorectal cancer several E545 mutations are subclonal and H1047 mutations clonal (Chang, 2016). The variation in hotspot distributions suggests different functions need to be targeted in different tumour types.
Different domain mutations in PIK3CA have been associated with differential signaling. Kinase domain mutations have higher activation of PI3K complexes and increased G-alpha signaling events. However, some components of the G-alpha signaling pathway such as RHO GTPase complexes have lower activity in the kinase domain mutants relative to other PIK3CA mutants. Additionally, helical domain mutations are associated with lower activation of pathways related to proliferation, such as FOXM1, MYC, and PLK1, in comparison to other PIK3CA mutants. The different patterns of pathway activation associated with mutations of specific PIK3CA domains suggest that kinase domain mutations are more strongly associated with proliferation activity, whereas helical domain mutations are more strongly associated with motility activities. These differences in function across domain specific mutations may drive the varying distribution of mutations seen across cancer types (Yau, 2014).

PIK3CA has a small CpG island in the gene promoter but no methylation has been described.