Correction: Data-driven prediction and design of bZIP coiled-coil interactions.

An error was introduced during the production process. A fixed-width font should have been used for the sequences in Table 3. The publisher apologizes for the error.

Table 3

Binding of designed peptides to their bZIP targets

Design	Target	K d at 37 °C (nM)	Designed sequence
XBP1-d1 1	XBP1	172 ± 65	EREAQLENRVAHLKEKNQELKAQNLHLKEALSEAQNRNQELKNDA
XBP1-d2	XBP1	≥ 5,000	AETDQLENRVKDLKKKNESLKEEKRQASNKYKALLTNNRSLKVKA
XBP1-d2* 2	XBP1	≥ 5,000 (269 at 23 °C)	AETDQLENRVKDLKKKNESLKEEKRQAKNKLKALLTNNRSLKVKA
JUN-d1 1	JUN	5.8 ± 0.7	SIAATLEKEEANLEKMNKKLAAEIESLLKEKDKLESVLNYHE
JUN-d2	JUN	1	QRALQLQKEKERLEKMNKKLAAEIESLLEERERLESVLNYHE
ATF3-d1	ATF3	~1,000	NDLARLENKAEELKVQNRILVDERKYLQREISELHDELAAHE
ATF3-d2	ATF3	564	NLVAQLEKKNEALKAENAALEIERIQLQDKIEELKYELAAIE
ATF3-d3	ATF3	≥ 5,000 (117 at 4 °C)	KDAASLENKKEELKVQNRILVDERKYLQMMNSELKDELAAHE
ATF4-d1 1	ATF4	9.3 ± 2.3	NQIKTLRTRLSKLRKDNLQLEKDIANLERKAKDLRAEKEQLEYEL
ATF5-d1 1	ATF5	1680 ± 1050	KRIAYLRQRIAELRNENHVLESRIQRMEKEKDALQQDRDHLEYEL
ATF5-d1 3	ATF4	4.9 ± 1.0	(same as above)

1 These four designs were chosen for further characterization.

2 A double mutant of XBP1-d2, in which serine at position 4e and tyrosine at position 5a were mutated to lysine and leucine, respectively.

3 Although ATF5 was used in the computational design, binding was tighter to ATF4, which shares 69.2% sequence identity with ATF5 in , , and positions.

Please view the corrected table here: