A biosynthetic dual-core cell computer April 23, 2019, 11:02

Con­trol­ling gene ex­pres­sion through gene switches based on a model bor­rowed from the dig­i­tal world has long been one of the pri­mary ob­jec­tives of syn­thetic bi­ol­ogy. The dig­i­tal tech­nique uses what are known as logic gates to process in­put sig­nals, cre­at­ing cir­cuits where, for ex­am­ple, out­put sig­nal C is pro­duced only when in­put sig­nals A and B are si­mul­ta­ne­ously pre­sent.

To date, biotech­nol­o­gists had at­tempted to build such dig­i­tal cir­cuits with the help of pro­tein gene switches in cells. How­ever, these had some se­ri­ous dis­ad­van­tages: they were not very flex­i­ble, could ac­cept only sim­ple pro­gram­ming, and were ca­pa­ble of pro­cess­ing just one in­put at a time, such as a spe­cific meta­bolic mol­e­cule. More com­plex com­pu­ta­tional processes in cells are thus pos­si­ble only un­der cer­tain con­di­tions, are un­re­li­able, and fre­quently fail.

Even in the dig­i­tal world, cir­cuits de­pend on a sin­gle in­put in the form of elec­trons. How­ever, such cir­cuits com­pen­sate for this with their speed, ex­e­cut­ing up to a bil­lion com­mands per sec­ond. Cells are slower in com­par­i­son, but can process up to 100,000 dif­fer­ent meta­bolic mol­e­cules per sec­ond as in­puts. And yet pre­vi­ous cell com­put­ers did not even come close to ex­haust­ing the enor­mous meta­bolic com­pu­ta­tional ca­pac­ity of a hu­man cell.

A CPU of bi­o­log­i­cal com­po­nents

A team of re­searchers led by Mar­tin Fusseneg­ger, Pro­fes­sor of Biotech­nol­ogy and Bio­engi­neer­ing at the De­part­ment of Biosys­tems Sci­ence and En­gi­neer­ing at ETH Zurich in Basel, have now found a way to use bi­o­log­i­cal com­po­nents to con­struct a flex­i­ble core proces­sor, or cen­tral pro­cess­ing unit (CPU), that ac­cepts dif­fer­ent kinds of pro­gram­ming. The proces­sor de­vel­oped by the ETH sci­en­tists is based on a mod­i­fied CRISPR-Cas9 sys­tem and ba­si­cally can work with as many in­puts as de­sired in the form of RNA mol­e­cules (known as guide RNA).

A spe­cial vari­ant of the Cas9 pro­tein forms the core of the proces­sor. In re­sponse to in­put de­liv­ered by guide RNA se­quences, the CPU reg­u­lates the ex­pres­sion of a par­tic­u­lar gene, which in turn makes a par­tic­u­lar pro­tein. With this ap­proach, re­searchers can pro­gram scal­able cir­cuits in hu­man cells – like dig­i­tal half adders, these con­sist of two in­puts and two out­puts and can add two sin­gle-digit bi­nary num­bers.

Pow­er­ful mul­ti­core data pro­cess­ing

The re­searchers took it a step fur­ther: they cre­ated a bi­o­log­i­cal dual-core proces­sor, sim­i­lar to those in the dig­i­tal world, by in­te­grat­ing two cores into a cell. To do so, they used CRISPR-Cas9 com­po­nents from two dif­fer­ent bac­te­ria. Fusseneg­ger was de­lighted with the re­sult, say­ing: “We have cre­ated the first cell com­puter with more than one core proces­sor.”

This bi­o­log­i­cal com­puter is not only ex­tremely small, but in the­ory can be scaled up to any con­ceiv­able size. “Imag­ine a mi­cro­tis­sue with bil­lions of cells, each equipped with its own dual-core proces­sor. Such ‘com­pu­ta­tional or­gans’ could the­o­ret­i­cally at­tain com­put­ing power that far out­strips that of a dig­i­tal su­per­com­puter – and us­ing just a frac­tion of the en­ergy,” Fusseneg­ger says.

Ap­pli­ca­tions in di­ag­nos­tics and treat­ment

A cell com­puter could be used to de­tect bi­o­log­i­cal sig­nals in the body, such as cer­tain meta­bolic prod­ucts or chem­i­cal mes­sen­gers, process them and re­spond to them ac­cord­ingly. With a prop­erly pro­grammed CPU, the cells could in­ter­pret two dif­fer­ent bio­mark­ers as in­put sig­nals. If only bio­marker A is pre­sent, then the bio­com­puter re­sponds by form­ing a di­ag­nos­tic mol­e­cule or a phar­ma­ceu­ti­cal sub­stance. If the bio­com­puter reg­is­ters only bio­marker B, then it trig­gers pro­duc­tion of a dif­fer­ent sub­stance. If both bio­mark­ers are pre­sent, that in­duces yet a third re­ac­tion. Such a sys­tem could find ap­pli­ca­tion in med­i­cine, for ex­am­ple in can­cer treat­ment.

“We could also in­te­grate feed­back,” Fusseneg­ger says. For ex­am­ple, if bio­marker B re­mains in the body for a longer pe­riod of time at a cer­tain con­cen­tra­tion, this could in­di­cate that the can­cer is metas­ta­sis­ing. The bio­com­puter would then pro­duce a chem­i­cal sub­stance that tar­gets those growths for treat­ment.

Mul­ti­core proces­sors pos­si­ble

“This cell com­puter may sound like a very rev­o­lu­tion­ary idea, but that’s not the case,” Fusseneg­ger em­pha­sises. He con­tin­ues: “The hu­man body it­self is a large com­puter. Its me­tab­o­lism has drawn on the com­put­ing power of tril­lions of cells since time im­memo­r­ial.” These cells con­tin­u­ally re­ceive in­for­ma­tion from the out­side world or from other cells, process the sig­nals and re­spond ac­cord­ingly – whether it be by emit­ting chem­i­cal mes­sen­gers or trig­ger­ing meta­bolic processes. “And in con­trast to a tech­ni­cal su­per­com­puter, this large com­puter needs just a slice of bread for en­ergy,” Fusseneg­ger points out.

His next goal is to in­te­grate a mul­ti­core com­puter struc­ture into a cell. “This would have even more com­put­ing power than the cur­rent dual core struc­ture,” he says.

By:  Peter Rüegg