There was an error in Development (2020) 147, dev191734 (doi:10.1242/dev.191734).
The authors have found that the cytokinin histidine kinase mutant lines used (hk5, hk6 and hk5 hk6) all contained, in addition to the stated hk5 and/or hk6 mutations, a further mutation in the HK4 gene. The mutation is a 1-bp insertion of an A residue at position 1179 in the genomic sequence (where the A in the start codon is +1):
WT HK4 sequence at CRISPR site: CGGTGGAGGATCGTGTTGCT
hk4-4 sequence at CRISPR site: CGGATGGAGGATCGTGTTGCT
The authors designated this allele hk4-4.
Once the authors discovered the problem, they sequenced all four HK genes. All versions of the lines used in the study were WT for HK3, homozygous mutant for the CRISPR-induced frameshift mutation in hk4, and had the expected hk5 and/or hk6 mutations. Therefore, the hk5 line reported in the paper is actually a hk4 hk5 double mutant, hk6 is a hk4 hk6 double mutant and hk5 hk6 is a hk4 hk5 hk6 triple mutant.
This does not substantially change the conclusions, except of course that the hk4 mutation probably also contributes to the phenotypes reported.