1. Volume 17, Issue 3, Pages 393-403 (February 2005)
Differential Targeting of Prosurvival Bcl-2 Proteins by Their BH3-Only Ligands Allows Complementary Apoptotic Function 
Lin Chen, Simon N. Willis, Andrew Wei, Brian J. Smith, Jamie I. Fletcher, Mark G. Hinds, Peter M. Colman, Catherine L. Day, Jerry M. Adams, David C.S. Huang 
Molecular Cell 
Volume 17, Issue 3, Pages (February 2005)
DOI: /j.molcel

2. Figure 1 Bim Binds Prosurvival Proteins Comparably
(A) Mcl-1 was injected onto sensorchips with mutant Bim4EBH3 (dotted line) or BimwtBH3 (dashed line) immobilized. To obtain the specific binding (solid line), the baseline response with 4EBH3 was subtracted from that with wtBH3. Irregular responses after 500 s are due to washing the chips after analyte dissociation.
(B) Prosurvival proteins bind Bim comparably. Sensorgrams showing the responses when prosurvival proteins were injected onto immobilized BimwtBH3.
(C) The binding constants of the prosurvival molecules for BimBH3 were determined from biosensor experiments as described previously (Hinds et al., 2003; Wilson-Annan et al., 2003).
Molecular Cell  , DOI: ( /j.molcel )

3. Figure 2 Competitive Binding Assays
(A) BH3 peptides used in the study. The immobilized peptides (top two rows) were derived from mouse (m) BimL. Bim4EBH3 has the four hydrophobic residues (h1–h4) required for interacting with the prosurvival proteins (see text) replaced by glutamates (E). Competitor peptides were derived from human proteins except those denoted “m.” Sequences were aligned as described (Huang and Strasser, 2000). Accession numbers of the sequences used are provided in the Experimental Procedures.
(B) Principle of solution competition assay. Occupancy of Bcl-xL by the competitor peptide (BikBH3 in this case) (1) competes with its binding to immobilized BimBH3 (2).
(C) Preincubation with a competitor BH3 peptide diminishes biosensor responses. Bcl-xL was preincubated with increasing concentrations of BikBH3 before the mixture was injected over a BimBH3 chip. The line (at 430 s) indicates the responses used for calculating the IC50.
(D) BikBH3 effectively competed with immobilized BimBH3 for Bcl-xL binding. The response (percentage of control; obtained from [C]) indicates the proportion of Bcl-xL that still binds the immobilized peptide in the presence of indicated concentrations of BikBH3, compared to Bcl-xL without BikBH3 (100%).
Molecular Cell  , DOI: ( /j.molcel )

4. Figure 3
Selective Interactions of BH3 Domains with Bcl-2-Like Proteins
(A) By using competitive binding assays (Figure 2), the IC50 (nM) for the indicated interactions were determined. The results shown are from representative experiments; the variation observed in multiple experiments was less than 2-fold (using different chips or protein batches).
(B) Relative binding affinities of interactions (A) were inversely plotted. The BH3 binding profiles of prosurvival proteins are shown.
(C) Biosensor responses to immobilized BadBH3 with recombinant prosurvival Bcl-2-like proteins (Bcl-2, Bcl-xL, Bcl-w, Mcl-1, and A1) or irrelevant controls (GST and LIF-receptor). Mcl-1 and A1 had no affinity for BadBH3, whereas Bcl-2, Bcl-xL, and Bcl-w bind BadBH3 avidly.
(D) Conversely, only Mcl-1 and A1 bound tightly to NoxaBH3.
Molecular Cell  , DOI: ( /j.molcel )

5. Figure 4
Bik, Bad, and Noxa Have Selective Prosurvival Targets in Mammalian Cells
Interactions between FLAG (FL)-tagged prosurvival proteins (human Bcl-2, human Bcl-xL, and mouse Mcl-1) and HA-tagged BH3-only proteins ([A], human Bim; [B], human Puma; [C], human Bik; [D] and [E], mouse Bad; [F and G], mouse Noxa) were tested by coimmunoprecipitation. Equivalent 35S-labeled lysates harvested from 293T cells were immunoprecipitated with antibodies to the HA, FL, or control (“C”) tags. Bim (A) or Puma (B) bound Bcl-2, Bcl-xL, and Mcl-1 well. In the top of (A), BimL was used instead of BimEL to discriminate Bim from Bcl-2 by size. (C) Bik preferentially bound Bcl-xL. (D) Bad bound Bcl-2 and Bcl-xL, but not Mcl-1, as confirmed by immunoblotting the same filters with the indicated antibodies (E). **Endogenous , associating with Bad, as confirmed by immunoblotting (E). (F) Noxa only bound Mcl-1, as confirmed by immunoblotting (G). *Degradation product of Mcl-1 incapable of binding.
Molecular Cell  , DOI: ( /j.molcel )

6. Figure 5
BH3-Only Proteins that Bind Selective Targets Have Weak Killing Activity
(A) The affinities of BH3-only peptides for prosurvival proteins (tabulated in Figure 3A) were inversely plotted to facilitate comparison of the patterns of BH3 binding.
(B) Potent killing of MEFs by Bim and Puma, but not Bmf, Bad, Bik, Hrk, or Noxa. Immortalized 3T9 MEFs were infected with retroviruses expressing only GFP (control) or one of the BH3-only proteins and GFP. The viability of the infected (GFP+ve) cells was determined by PI exclusion 24 hr after infection. The histograms represent means ± one SD of at least three experiments.
(C) Interactions between FL-tagged prosurvival proteins (Bcl-xL and Mcl-1) and BimS (or its variants) were tested by coimmunoprecipitation. Equivalent lysates from transfected 293T cells were immunoprecipitated with antibodies to Bim, FL tag, or a control irrelevant antigen (“C”). The filter was probed with a rat monoclonal anti-FL antibody. *Degradation product of Mcl-1.
(D) BimS variants that have restricted binding to the prosurvival proteins are weak killers. Viability of MEFs infected with retroviruses encoding the indicated proteins (GFP+ve) was analyzed 24 hr after infection. The histograms represent means ± one SD of at least three experiments.
(E) Long-term survival of MEFs infected with BH3 expressing retroviruses. One hundred retrovirally infected cells were plated and the absolute number of GFP+ve colonies formed after 6 days scored. No colonies were obtained after infection with BimS (†), whereas BimS 4E did not affect long-term viability. Bik, Noxa, BimS BadBH3, or BimS NoxaBH3 had modest effects. The data represent the average number of GFP+ve colonies formed ± one SD from at least three experiments.
Molecular Cell  , DOI: ( /j.molcel )

7. Figure 6 Cooperation between Different Classes of BH3-Only Proteins
(A) Based on our binding data, a model proposed to explain the weak killing activity of certain BH3-only proteins that have selective targets (see text).
(B) Cooperation between proapoptotic BH3-only proteins. MEFs were infected with retroviruses coexpressing a BH3-only protein (BimS, Bik, Noxa, or Noxa3E) and GFP or a BH3-only protein and GFP-tagged-BimS, -BimS BadBH3, or -BimS NoxaBH3. Cell viability was scored 24 hr after infection; data represent means ± one SD of at least three experiments.
Molecular Cell  , DOI: ( /j.molcel )

8. Figure 7 A Less Selective Noxa Mutant Is a Potent Killer
(A) Interactions between the α-helical BimBH3 region (light blue; key residues labeled in red; the numbering refers to mouse BimL) with the target groove of Bcl-xL (key residues labeled in black) (Liu et al., 2003) (see text).
(B) Alignment of the core BH3 regions of human Bim (BimL) and human Noxa. Mutated Noxa residues are boxed in red.
(C) Increased binding of Noxa mutants to Bcl-xL and Bcl-w. Wild-type or mutant Noxa peptides were tested in solution competition assays for their capacity to bind Bcl-2, Bcl-xL, Bcl-w, or Mcl-1. The histograms show the IC50 (nM) for each interaction. The crosses symbolize that IC50 > 100 μM.
(D) Noxa m3 binds both Bcl-xL and Mcl-1. Interactions between Bcl-xL or Mcl-1 and BimS NoxaBH3 m3 were tested by coimmunoprecipitation as described in Figure 4. *Mcl-1degradation product.
(E) NoxaBH3 m3 is a potent killer. Survival of MEFs 24 hr after infection with the indicated retroviruses.
(F) Dose-dependent killing by NoxaBH3 m3. Viability of MEFs infected with NoxaBH3 or NoxaBH3 m3 sorted for low, medium or high GFP expression; data in (E) and (F) represent means ± one SD of at least three experiments.
Molecular Cell  , DOI: ( /j.molcel )