3.2 Effects of antibody affinity on yeast/mammalian cell
interactions
We sought to compare the impact of antibody binding affinity on
yeast/mammalian cell interactions in the biofloating and biopanning
platforms. For biofloating characterization, yeast cells expressing the
atezolizumab, D12, A1, and nivolumab (an anti-PD-1 antibody serving as a
negative control) scFvs were co-incubated with PD-L1+CHO-K1 cells in suspension with a yeast:mammalian cell ratio of 10:1.
The co-incubation period was varied from 0 to 180 min to gauge the
kinetics of this system under different affinity conditions. We found
that yeast displaying the high affinity atezolizumab and medium affinity
D12 scFvs were fully bound to PD-L1+ CHO-K1 cells
virtually instantaneously (Figures 2a and 2c). However, the low affinity
clone A1 showed a time-dependent increase in mammalian cell binding, and
did not achieve the same degree of binding as the high affinity or
medium affinity scFvs, even after 180 min. As anticipated, no binding of
atezolizumab, D12, or A1 scFvs was detected on PD-L1-CHO-K1 cells via biofloating (Figure S2a). For biopanning
characterization, scFvs were co-incubated with the
PD-L1+ CHO-K1 monolayers. We found that yeast
displaying the high and medium affinity clones showed detectable binding
to PD-L1+ CHO-K1 cells within 60 min, whereas yeast
displaying the low affinity clone did not show detectable binding to
mammalian cells within 180 min (Figures 2b and 2d). As expected, no
binding of atezolizumab, D12, or A1 scFvs was detected on
PD-L1- CHO-K1 cells via biopanning (Figure S2b). In
contrast with their instantaneous saturation using the biofloating
platform, interaction between the high and medium affinity clones and
the mammalian cells increased over time, demonstrating the stronger
kinetic dependence of yeast/mammalian cell binding for the biopanning
versus the biofloating platform. Moreover, detection of interaction
between yeast displaying the low affinity clone and mammalian cells in
the biofloating but not the biopanning setup indicates the superior
sensitivity of the former.
We next sought to compare the impact of antibody binding affinity on the
optimal yeast:mammalian cell co-incubation ratio in both the biofloating
and biopanning platforms. To this end, yeast displaying atezolizumab,
D12, A1, and nivolumab (negative control) scFvs were incubated with
PD-L1+ CHO-K1 cells at various yeast:mammalian cell
ratios while keeping the incubation time constant. For biofloating
experiments, binding was detectable for ratios as low as 0.57 for the
high, medium, and low affinity clones (Figure 3a). The potency of
binding corresponded with scFv affinity, with the high affinity clone
requiring the lowest yeast:mammalian cell ratio to achieve saturation
(Figure 3c). Based on these findings, a 10:1 yeast:mammalian cell ratio
was determined to be sufficient for biofloating studies. Analogous
biopanning studies demonstrated that >10-fold higher
yeast:mammalian cell ratios were required to achieve saturation, as
compared to biofloating (Figures 3b and 3d). Binding of yeast displaying
the low affinity scFv to mammalian cells was barely detectable, even at
the highest yeast:mammalian cell ratio, again supporting the enhanced
sensitivity of binding detection for the biofloating platform. Our
results also confirm that the 40:1 yeast:mammalian cell ratio previously
implemented for biopanning studies18 is sufficient to
reach saturation for high affinity scFv clones.