14.5 De novo Organelle Construction and Future Directions 295the N‐terminus of carboxypeptidase Y, to deliver the Batis maritima methyl
halide transferase to the vacuole and increase productivities for methyl iodide
10‐fold [114]. Further, taking advantage of the fact that SAM levels can be
increased by altering media conditions, methyl iodide production was
increased an additional fivefold by stimulating SAM production [115].
14.5 De novo Organelle Construction and Future Directions
Directions
In other cases, it may be advantageous to start with cellular structures that can
be repurposed to a larger extent and, perhaps, to create organelle function
de novo. The simplest version of this idea is to completely hijack an existing
organelle with less essential function. For example, peroxisomes are oxidative
organelles that sequester the toxic reactions methyltrophy and/or very‐long‐
chain fatty acid catabolism, but are not required for cellular viability under most
conditions [116]. As such, they are an intriguing target for engineering. Under
standing peroxisome biogenesis is still a work in progress, but proteomics exper
iments indicate the peroxisome contains about 10× fewer proteins than the
mitochondrion, lending credence to the idea of simplicity [117]. Importantly,
there are also well‐defined targeting signals to both the matrix of the peroxisome
and the membrane [118, 119]. The biosynthetic capability of the peroxisome
has been exploited to improved production of biofuels such as fatty alcohols
[120, 121] and alkanes [121].
A more ambitious area of research is to construct an entire organelle‐like
structure de novo. From a materials science perspective, organelles are formed
when a set of molecular building blocks spontaneously self‐organize, through
molecular interactions, into complex patterns [24]. De novo design therefore
requires identifying and engineering self‐organizing building blocks. Additional
properties that emerge from this process are organelle size/shape and copy
number. These too must be accounted for. Recent work from Lim and col
leagues demonstrates how this may be possible [122]. By leveraging the various
lipid binding and lipid synthesis/degradation domains from the phosphati
dylinositol signaling pathway, coupled with positive and negative feedback,
Chau et al. were able to create pole‐localized lipid microdomains. Given the
large toolkit of phosphatidylinositol‐binding domains, these microdomains
could serve as the initial scaffold for generating more complex structures. In an
orthogonal approach, Eriksson and colleagues demonstrated that overexpres
sion of an integral membrane lipid glycosyltransferase yields massive vesicle
formation in E. coli [123]. Combining these approaches may enable the crea
tion of targetable distinct lipid‐bound structures with controllable size and
copy number, although this remains a tremendous challenge. However, such a
synthetic organelle may also help to shed light on the natural organelle biogen
esis process [24].
Finally, it may be revealing to reflect upon even more complex engineer
ing challenges. Biology clearly takes advantage of compartmentalization far
beyond a single genome. For example, the sea slug Elysia chlorotica carries out