Taq polymerase has a speed of 60 – 100 nucleotides per second (nuc/sec). The usual recommendation is to run the extension step in a PCR assay at 1 min/kb for Taq polymerase (due to several factors), which gave me the impression that it is a slow polymerase. Meanwhile, NEB Q5 and NEB Phusion have a recommended extension speed of 15 – 30 sec/kb, about 2 – 4x faster than Taq. However, speed comparison between DNA polymerases is not the point of this essay.
Assuming an average speed of 80 nuc/sec, I could not get a sense how fast that means. I then set out on a small adventure to conceptualize this. The substrate for Taq polymerase, as we all know, is single-stranded DNA (ssDNA), of which Taq catalyzes the templated addition of nucleotides to yield double-stranded DNA (dsDNA). DNA double helix is about 2 nanometer (nm) in width, and the length of a single base pair is 0.34 nm. Taq, a globular protein with the molecular weight of 94 kDa (832 residues) has an approximate size of 5 – 7 nm, of which we will assume 6 nm to simplify the calculation.
At 80 nuc/sec, Taq travels 27.2 nm in a second. This is about 4.53x of Taq’s diameter (speed:diameter ratio).
Meanwhile, the average running speed of a healthy human adult is 8 meter/sec. At an average adult male height of 1.75 m, that is 4.57x speed:height ratio. Usain Bolt (1.96 meter) can hit 10.44 meter/sec, yielding a 5.33x speed:height ratio. Taq is a touch slower than the average adult male, and definitely cannot compete against Usain Bolt.
On the other than, human RNA polymerase II (RNAPII) runs much slower at 20 – 30 nuc/sec when synthesizing RNA from DNA template, giving it a 0.4 – 0.6x speed:diameter ratio (15 - 20 nm in size; 550 kDa a 12-subunit enzyme complex); nearly 10x slower than Taq. However, RNAPII has wildly higher processivity; it could keep transcribing RNA for 1 – 2 megabases per binding event. Taq polymerase could manage 5 – 10 kilobase per binding event, with some estimates being much lower and therefore unsuitable for long-range DNA synthesis. RNAPII’s processivity is a curious thing, as most human genes are around 20 – 50 kilobases. That means, in a single binding event, RNAPII synthesizes RNA enough for ~43 genes, assuming all genes are equal in characteristics. It could do so in 16.7 hours, which is slow. If Taq were the RNAPII, it could have done it in 5.2 hours.