Mechanical diversity and folding intermediates of parallel-stranded G-quadruplexes with a bulge

Abstract A significant number of sequences in the human genome form noncanonical G-quadruplexes (G4s) with bulges or a guanine vacancy. Here, we systematically characterized the mechanical stability of parallel-stranded G4s with a one to seven nucleotides bulge at various positions. Our results show that G4-forming sequences with a bulge form multiple conformations, including fully-folded G4 with high mechanical stability (unfolding forces > 40 pN), partially-folded intermediates (unfolding forces < 40 pN). The folding probability and folded populations strongly depend on the positions and lengths of the bulge. By combining a single-molecule unfolding assay, dimethyl sulfate (DMS) footprinting, and a guanine-peptide conjugate that selectively stabilizes guanine-vacancy-bearing G-quadruplexes (GVBQs), we identified that GVBQs are the major intermediates of G4s with a bulge near the 5′ or 3′ ends. The existence of multiple structures may induce different regulatory functions in many biological processes. This study also demonstrates a new strategy for selectively stabilizing the intermediates of bulged G4s to modulate their functions.

NMR structural studies showed that the HIV-1 integrase inhibitor T30177 formed a parallel-stranded G4 structure with one nucleotide (nt) thymine bulge that connects two adjacent guanines of one column (9). All 12 guanines in this sequence participated in G-tetrad formation. A systematic analysis revealed that the bulge can form in 8 different positions between the 8 successive guanine pairs with sizes of 1 to 7 nt, and all these sequences form parallel-stranded G4s (10). Circular dichroism (CD) and NMR melting curve analysis revealed that the bulge reduced the thermodynamic stability of G4s compared with the original T30695 sequence, which does not have a bulge. The stability of the bulged G4s depends on the size and the position of the bulges, in which the TB-1 sequence (Table 1) has the high-2 Nucleic Acids Research, 2021 est melting temperature (T m ) compared to other sequences (TB-2 to TB-8) (10). When the bulge length increased from 1 nt (TB-1) to 7 nt (T7B-1), T m decreased >37 • C in 60 mM K + buffer (10). However, G4-forming sequences often exhibit complex folding energy landscapes and diverse folding conformations (i.e., long-lived intermediates) (25,26). These ensemble methods encounter difficulty in resolving the stability of multiple coexisting G4 structures.
Single-molecule force spectroscopy techniques have become useful tools for studying the mechanical diversity of G4 structures (27)(28)(29)(30)(31)(32)(33)(34)(35)(36)(37). The Mao lab reported that the ILPR promoter sequence forms two different G4 conformations with distinct mechanical stability (28). Based on the unfolding force distributions and unfolding step size distributions, the Mao lab identified multiple intermediates of hTERT promoter G4s (29) and human telomeric G4s (30). By combining optical tweezers and fluorescence resonance energy transfer (FRET), the Ha lab observed that telomere G4-forming sequences can form more than six stable secondary structures with distinct mechanical stability (34,35). Hence, single-molecule measurements of mechanical stability are well suited for studies seeking to characterize the polymorphism of G4 formation. In addition, DNA G4s are subjected to various mechanical modifications during transcription and replication. DNA or RNA polymerases and helicases function as motor proteins tracking along with DNA and exerting forces on G4 structures (38). Therefore, the measurements of the mechanical stability of G4s provide insights for understanding the interactions between G4s and the motor proteins (39). However, the mechanical stability and conformational diversity of noncanonical G4s remain poorly understood compared with canonical G4s.
Herein, we systematically investigated the mechanical stability of noncanonical parallel-stranded G4s containing a one to seven nt bulge at different positions (Table 1) and 12  GVBQs (Supplementary Table S1) by using single-molecule magnetic tweezers. To identify the potent intermediates of G4s with a bulge located near the 5 or 3 end, we used a guanine-RHAU23 peptide conjugate (GRPC), which selectively stabilizes the GVBQs (24). This bifunctional GRPC was composed of a guanine base and a G4 binding domain (RHAU23) from the RHAU helicase (24).

Oligonucleotides, peptide, DNA sample preparation
Oligonucleotides were purchased from Sangon Biotech Co., Ltd (China) and Genewiz, Inc (China). GRPC peptide (a guanine moiety connect to a 23 amino acids peptideHPGH LKGREIGMWYAKKQGQKNK) was purchased from SBS Genentech (China) as previously described (24). The DNA constructs containing the G4-forming sequence for single-molecule experiments were prepared as previously described (40,41).

CD spectroscopy
The CD spectra and CD melting curves were collected on a Jasco-810 spectropolarimeter using a 1-mm optical path length quartz cuvette. Scans were performed in a range of 220 to 320 nm for five times at room temperature with a speed of 500 nm/min. The oligonucleotides were dissolved in 10 mM Tris-HCl (pH 8.0) buffer containing 100 mM and 20 mM KCl for CD spectra and CD melting curve measurements, respectively. The oligonucleotides were heated at 95 • C for 5 min and slowly cooled to room temperature. The melting experiments were performed at several temperature points during heating from 20 • C to 85 • C (1.0 • C/min) and the temperature was equilibrated for 5 min before recording. The melting curve profile was fitted by a sigmoid function utilizing the absorbance at 265 nm.

Dimethyl sulfate (DMS) footprinting
Fluorescein (5 -FAM) labeled oligonucleotides (0.1 M) dissolved in 10 mM Tris-HCl (pH 8.0) and 0.5 M EDTA buffer were heated at 95 • C for 5 min and subsequently cooled to room temperature. Next, 100 mM LiCl, 100 mM KCl or 0.5 M GRPC/100 mM KCl was added to the oligonucleotides, and the mixtures were immediately treated with 5% DMS reagent (Innochem, China) for 2 min on ice. Reactions were terminated by adding 80 l stop buffer (3 M sodium acetate, 0.1 M ␤-mercaptoethanol, and 1 mg/ml spermidine DNA). After chloroform extraction and ethanol precipitation, DNA was cleaved by the addition of 10%, vol/vol piperidine (Sinopharm Chemical Reagent Co.,Ltd, China) and incubated at 90 • C for 30 min. DNA was purified again by chloroform extraction and ethanol precipitation and dissolved in 80% (vol/vol) deionized formamide (Sangon Biotech Co., Ltd, China). The DNA samples were boiled at 95 • C for 5 min and subsequently cooled on ice for 15 min before being loaded onto a 20% denaturing polyacrylamide gel. DNA fragments were visualized by iBright 1500 (Thermo Fisher Scientific, USA) and digitized by ImageQuant 5.2 software.

Mechanical stability of TB-1 to TB-8 G4s bearing a 1 nt bulge depends on bulge positions
We first analyzed G4s with a 1 nt thymine bulge between two adjacent guanine residues at 8 different positions (Table 1, TB-1 to TB-8). A previous NMR study (10) and CD spectra measurements showed that the TB-1 to TB-8 and T30695 sequences form parallel-stranded G4 structures in the presence of 100 mM KCl ( Figure 1A). To characterize the mechanical stability, we used singlemolecule magnetic tweezers ( Figure 1B) to measure the unfolding force distribution based on a force-ramp procedure (Supplementary Figure S1), as described previously (40,41). The unfolding force distributions of the TB-1 to TB-8 and T30695 sequences exhibit multiple unfolding force peaks, suggesting that these sequences form multiple conformations ( Figure 1C-D). The TB-2 to TB-7 sequences show a major peak centered at ∼55 pN and a minor peak centered at ∼16 pN ( Figure 1D, Supplementary Table S2), which are similar to the T30695. The average unfolding step sizes of the major form of TB-2 to TB-7 are consistent with the total number of nucleotides in fully-folded bulged G4s (Supplementary Figure S2A-B). This result suggests that the TB-2 to TB-7 sequences primarily form fully-folded parallelstranded G4s with high mechanical stability. By fitting the unfolding force distributions with Bell's model (Supplementary, data analysis) (44,45), we obtained the zero-force unfolding rates k unfold for the fully-folded bulged G4s (TB-2 to TB-7), which were in the range of 10 -5 to 10 -7 s -1 (Supplementary Figure S2C). Previously, we showed that T30695 and other canonical parallel-stranded G4s have a major un-folding force peak at 40-60 pN, which is associated with zero-force unfolding rates of 10 -5 to 10 -7 s -1 (37). This result suggests that the characteristics of high mechanical stability and slow unfolding rates can be applied to noncanonical G4s with a 1 nt bulge.
In contrast to TB-2 to TB-7 sequences, the TB-1 sequence, which has a bulge near the 5 end, reveals large fractions of unfolding events at < 40 pN. By fitting to Bell's model, we obtained three unfolding force peaks for the TB-1 sequence (∼22 pN, ∼36 pN and ∼47 pN) and the fractions of each peak (62%, 25% and 7%). The average unfolding step sizes (16 nt) of the mechanically stable state is consistent with the total number of nucleotides of the fullyfolded TB-1 G4 structure (Supplementary Figure S3A). The unfolding step sizes of the two less stable states were determined to be 14 nt and 13 nt, suggesting the formation of partially-folded G4s. When the time for refolding at 1 pN was reduced from more than 30 s to 1 s, the TB-1 sequence formed only two less stable states, suggesting that the less stable states were kinetically favored states (Supplementary Figure S3). The TB-8 sequence shows two unfolding force peaks centered at ∼31 pN (fraction = 33%) and ∼46 pN (fraction = 63%) with unfolding step sizes of 14 nt and 16 nt, respectively, indicating the formation of partially-folded and fully-folded G4s (Supplementary Figure S3). We next measured the time-evolution folding probability p fold (t) of TB-1 to TB-8 sequences. The DNA constructs were held at low force of 1 pN for various holding time intervals t and then unfolded in a force-ramp stretching cycle. The folded states reveal an unfolding event, while the unfolded or misfolded states reveal single-stranded DNA (ssDNA) like mechanical behavior in a subsequent stretch-  (5) TB-4 n = 148 (6) TB-3 n = 219 (12) TB-5 n = 249 (14) TB-6 n = 230 (11) TB-7 n = 128 (14) TB-8 n = 257 (19) TB-1 n = 245 (8) TB-2 n = 215 (9) ing cycle. The time-evolution folding probability of each sequence was fitted with a single-exponential function, p fold (t) = p st [1 -exp(-k fold t)], where p st is the steady-state folding probability and k fold is the apparent folding rate of all folded species ( Figure 1E). Based on the unfolding forces, we also estimated the p st values of fully-folded G4s (unfolding forces > 40 pN, Figure 1F red column) and partially-folded G4s (unfolding forces < 40 pN, Figure 1F, gray column). The TB-1 and TB-8 sequences exhibit relatively large fractions of partially-folded G4s (87% and 33%) and high p st (94% and 96%). In contrast, the TB-2 to TB-7 sequences primarily form fully-folded G4 and show p st in the range of 58-79% ( Figure 1F), suggesting that a 1 nt bulge at the middle increased the unfolded or misfolded fractions. The apparent folding rate of the TB-1 (k fold = 0.11 ± 0.03 s -1 ) sequence was one order of magnitude higher than that of the TB-2 to TB-7 sequences (k fold = ∼0.01-0.03 s -1 ) ( Figure 1G), which also supports that the TB-1 sequence folds into kinetically preferred partially-folded G4s.

Sequences with a bulge near the 5 or 3 end form intermediates with a guanine-vacancy
To analyze the effects of bulge length, we measured the mechanical stability of G4-forming sequences bearing a different number of thymines between successive guanine pairs near the 5 end (T3B-1, T5B-1, and T7B-1) and 3 end (T3B-8, T5B-8 and T7B-8) ( Table 1). As the bulge length increases, all the sequences form two less stable states with unfolding forces of ∼20 and ∼36 pN (Figure 2A, gray column, Supplementary Table S2). Notably, the unfolding step sizes of G4s with different lengths of a bulge near the 5 end (TB-1, T3B-1, T5B-1 and T7B-1) and 3 end (TB-8, T3B-8, T5B-8 and T7B-8) are all ∼14 nt, which means that they are shorter than the fully-folded G4 structures ( Figure 2B, Supplementary Figure S4). The ∼14 nt unfolding step sizes are consistent with partially-folded G4 structures with a peeled guanine, suggesting that the guanine at the 5 or 3 end does not participate in the G-tetrad core (also called GVBQs). In addition, both the folding probability (p st ∼90%, Supple-

B
Step sizes (nt) mentary Figure S5A and B) and folding rates (k fold ∼0.1 s -1 , Supplementary Figure S5C) of the TB-1 and TB-8 groups were independent of the bulge length, suggesting that these sequences with a bulge near their 5 or 3 ends were kinetically trapped in intermediate states.
To determine whether the intermediates of TB-1 and T3B-1 are GVBQs, we added GRPC to selectively target GVBQs ( Figure 2C) (24). In the presence of a saturated concentration of 0.5 M GRPC, TB-1 and T3B-1 show two unfolding force peaks at ∼33 pN, ∼53 pN and at ∼40 pN, ∼55 pN, respectively (Figure 2A, red column). The ∼20 pN major unfolding force peaks of TB-1 and T3B-1 disappear, and a new unfolding force peak at ∼54 pN appears, suggesting that the GRPC binds the intermediate of TB-1 and T3B-1 and increases the unfolding forces from ∼20 to ∼54 pN. The complementation of the peeled 5 end guanine of the intermediates by GRPC results in similar mechanical stability as fully-folded parallel-stranded G4. Similar stabilization effects by GRPC were obtained for the TB-8 and T3B-8 G4s, suggesting that GRPC can also bind and stabilize 3 end peeled GVBQs. It should be noted that there is more than one intermediate for TB-1 and T3B-1 and the ∼36 pN peak of TB-1 and T3B-1 is retained, suggesting that GRPC cannot stabilize this intermediate. The GRPC also reduced the folding probability p fold probably due to the interactions between GRPC and other states (Supplementary Figure S5D).
To further confirm that GRPC stabilizes intermediate G4 through filling in guanine at the 5 or 3 end, we performed DMS footprinting experiments ( Figure 2D). The 5 gua-nine of T3B-1 shows stronger cleavage in the presence of 100 mM K + and 0.5 M GRPC (K + /GRPC) compared with 100 mM K + only. This result suggests that the binding of GRPC to partially-folded T3B-1 blocks the conversion from partially-folded GVBQs to fully-folded G4s, thereby resulting in stronger cleavage of 5 guanine. The TB-1 sequence does not exhibit stronger cleavage of 5 guanine in the presence of K + /GRPC, probably because TB-1 has a faster rate than T3B-1 to fold into fully-folded G4s (Supplementary Figure S6). TB-8 and T3B-8 sequences also revealed stronger cleavage of the peeled 3 guanine in the presence of K + /GRPC than in the presence of only K + (Supplementary Figure S6). Taken together, these results indicate that G4-forming sequences with a bulge near the 5 or 3 end kinetically access partially-folded GVBQs, which can be selectively stabilized by GRPC.

Mechanical stability of GVBQs depends on the positions of missing guanine
To confirm that the GVBQs have weaker mechanical stability than fully-folded G4s, we also analyzed 12 G4-forming sequences with a guanine-to-thymine substitution (named G1-T to G12-T) (Supplementary Table S1). CD spectra (Supplementary Figure S7A) and a previous NMR study (13) suggested that these sequences form parallel-stranded GVBQs. The unfolding force distributions of 12 G-to-T substitution mutants measured using the same forceramp procedure show that the major unfolding force peaks are < 40 pN ( Figure 3A, Supplementary Table S2), sug-   gesting that GVBQs have lower mechanical stability than fully-folded parallel-stranded G4s. The presence of multiple peaks for G1-T, G2-T, G7-T, G10-T, G11-T and G12-T suggests the formation of multiple structures with different mechanical stabilities. After 0.5 M GRPC was added to the flow chamber, the major unfolding force peaks for G1-T, G2-T, and G3-T sequences were shifted from ∼25, ∼17 and ∼22 pN to ∼53, ∼42 and ∼54 pN, respectively ( Figure 3A). These results suggest that the binding of GRPC to GVBQs stabilized it and resulted in mechanical stability as high as fully-folded parallel-stranded G4, which is consistent with the results obtained from TB-1 and T3B-1 bulged G4s. An interesting finding in our single-molecule measurements is that the folding probabilities of G4s with a Gto-T substitution strongly depend on the positions of the missing guanine. Based on the positions of the missing guanine, we grouped the sequences as top-tetrad, middletetrad, and bottom-tetrad mutants that contain G-tract 5 TGG, 5 GTG and 5 GGT, respectively. Both top-tetrad and bottom-tetrad mutants show high folding probability and rapid folding rates (top-tetrad p st ∼90%, k fold ∼0.09 s -1 ; bottom-tetrad p st ∼85%, k fold ∼0.07 s -1 ) compared with the middle-tetrad mutants (p st ∼24%, k fold ∼0.02 s -1 ) (Figure 3B, Supplementary Figure S7B-D). Addition of 0.5 M GRPC increased the folding probability p fold of middle tetrad G2-T but reduced the p fold of G1-T and G3-T though the reason remains unclear ( Supplementary Figure S7E). The low folding probability of middle-tetrad mutants suggests that the abolished consecutive guanine pairs in one column of the G-tetrad core strongly hindered stable G4s formation. This finding is consistent with a previous single-molecule study which showed that a middle tetrad G-to-T mutant of telomeric G4s displayed unstructured ssDNA-like behavior in 100 mM K + (35). The positiondependent folding probability is correlated with the thermodynamic stability (T m s) of G4s ( Figure 3C). The top-tetrad and bottom-tetrad mutants show a considerably higher T m (top-tetrad, T m ∼48 • C, bottom-tetrad T m ∼44 • C) than the middle-tetrad mutants (T m ∼36 • C) in 20 mM KCl (Figure 3C, Supplementary Figure S7F), suggesting that the middle-tetrad mutants have much lower thermodynamic stability. The lower melting temperature of guanine substitution in the middle-tetrad have also been observed for G-to-oxo G and G-to-A substitutions of the telomeric G4forming sequence (46,47), suggesting that this is a general feature of guanine substitution mutants.

Mechanical stability and bulge-length dependence of G4s with a ≥ 2 nt bulge in the middle
To evaluate the effects of bulge positions and bulge lengths in the middle of the sequences, we measured the mechanical stability of T2B-2 to T2B-7 sequences with a 2 nt bulge ( Table 1). CD spectra show that all these sequences fold into parallel-stranded G4 structures ( Supplementary Figure S8). Based on the positions of the bulge, we grouped the sequences with a bulge at position 2, 4, 6 (T2B-2, T2B-4, and T2B-6) as lower bulged G4s and the sequences with a bulge at position 3, 5, 7 (T2B-3, T2B-5, and T2B-7) as upper bulged G4s. The unfolding forces and unfolding step sizes suggest that the lower bulged G4s mainly form fully-folded G4s with high mechanical stability (Supplementary Figure  S8). Comparing with G4s with a 1 nt lower bulge (Figure 1), the increased 2 nt bulge length reduced the folding probabil-  Figure S8). In contrast to lower bulged G4s, the upper bulged G4s exhibit much broader unfolding force distributions between 5 and 60 pN which suggests the formation of multiple structures with distinct mechanical stability. The differences between upper and lower bulged G4-forming sequences could depend on the G4s topology.
Next, we increased the length of thymines at position 2 (TxB-2: T2B-2, T3B-2, T5B-2, T7B-2) and position 3 (TxB-3: T2B-3, T3B-3, T5B-3, T7B-3) ( Table 1). The T2B-2, T3B-2 sequences show a major peak at ∼56 pN ( Figure 4A, Supplementary Table S2), while other sequences show broad unfolding force distributions with multiple peaks ( Figure  4B). The structures with unfolding forces > 40 pN reveal unfolding step sizes of fully-folded G4s, while the structures with unfolding forces < 40 pN reveal smaller unfolding step sizes indicating the formation of partially-folded G4s ( Figure 4A-B). To test whether the partially-folded structures contain GVBQ structures, we measured the unfolding forces of T2B-3 in the presence of 0.5 M GRPC ( Figure  4C). The < 40 pN species of T2B-3 were retained suggesting that these partially-folded structures cannot be stabilized by GRPC, which implies the formation of other conformations. By analyzing the time-dependent folding probability of TxB-2 and TxB-3 sequences, we obtained the steady-state folding probability p st and folding rates k fold (Supplementary Figure S9). Due to the low folding probability, the folding probability p fold (t) of the T7B-2 and T7B-3 measured after holding the DNA at low force for 300 s were used for comparisons ( Figure 4D). The overall p st of TxB-2 and TxB-3 decrease from 69% (TB-2) and 79% (TB-3) to 9% (T7B-2) and 10% (T7B-3), when the bulge length increases from 1 to 7 nt. For the TxB-3 group, the fraction of fullyfolded G4s was lower than 15% when the bulge was > 2 nt ( Figure 4D, red column). This result indicates that as the bulge became longer, the middle bulge significantly reduced the folding probability of fully-folded and stable partiallyfolded G4s.

DISCUSSION
Mechanical diversity and partially-folded G4s such as triplexes have been well documented for canonical G4s (29,30,35,37,48). However, the polymorphism and the effects of a bulge on the mechanical stability of noncanonical G4s have not been investigated before. We systematically analyzed the mechanical stability of bulged G4s and our results suggest that G4-forming sequences with a bulge can form multiple structures, including fully-folded G4s (unfolding forces > 40 pN), partially-folded intermediates (unfolding forces < 40 pN) ( Figure 5). This result is in contrast to the findings of previous NMR studies, which revealed that only fully-folded TB-1 G4s with all 12 guanines participate in the formation of the G-tetrad core (9,10). This discrepancy may be observed because the G4s prepared in NMR measurements undergo thermal annealing processes over several hours, while the G4s refold in the force-ramp procedure is a seconds-to-minutes time-scale. Many studies have revealed that reaching the equilibrium state (thermodynamic minima) of G4s requires a long time period (up to days) (49). The kinetically preferred intermediates may be biologically relevant states, as many biologically relevant processes, for example, transcription, occur on a secondsto-minutes time scale. The diverse intermediates with different mechanical stability may function as different barriers to the progression of motor proteins, for example, helicases or polymerases (38).
The identification of GVBQs as major intermediates of end bulged G4s is important for understanding the folding pathways of bulged G4s. The formation of GVBQs or strand-slipped G4s as a folding intermediate has been proposed through molecular dynamics simulations (50,51). However, there was no direct evidence that G4s can form via GVBQ intermediates. Based on the unfolding step sizes and unfolding forces, we established for the first time that GVBQs are major intermediates of G4s with a bulge near the 5 or 3 end while the 5 or 3 end guanine flips out and does not participate in the formation of G4. With increasing time, the GVBQs slowly convert to fully-folded bulged G4s (Supplementary Figure S3). Figure 5 upper panel shows two potent GVBQs structures while the first GVBQs can transition to the fully-folded G4s by simply docking of the 5 guanine and the second GVBQs must undergo strand shift before the 5 guanine can dock into the structure. The strand shift may take a long time thus explaining the slow conversions from the GVBQs to fully-folded G4s. Our data show extreme mechanical diversity of middle bulged G4s (T2B-3), suggesting the formation of more than one partiallyfolded intermediates. We also provide a new strategy to modulate G4s functions by stabilizing kinetically preferred GVBQ intermediates. By using GRPC specifically to stabilize GVBQs, we demonstrated that selective filling in of the guanine of GRPC into the G-vacancy can achieve high mechanical stability as fully-folded G4s, thus represent promising targets for drug design.
The systematic analysis of the effects of bulge lengths and positions on the mechanical stability of G4s may help to establish rules to predict the mechanical stability and folding probability of noncanonical G4s. Our previous measurements on canonical parallel-stranded G4s suggest that high mechanical stability is prevalent in fully-folded parallelstranded G4s (37,40,52). The current study further demonstrates that parallel-stranded G4s with a bulge (fully-folded states) also exhibit high mechanical stability (unfolding forces > 40 pN), while parallel-stranded GVBQs exhibit low mechanical stability. In addition to previous reported melting temperatures (10), we show that folding probability, apparent folding rates, and folded populations depend on both bulge positions and bulge lengths ( Figure 5). A bulge reduced the thermal stability of G4s probably because the bulge length dependent entropy cost of closing a loop or bulge in G4 structures (53,54). When a bulge is located near the 5 or 3 end, the sequences primarily form kinetically preferred GVBQs. The entropy cost for closing the bulge at the 5 or 3 end prevents the convention of GVBQs to fully-folded G4s. When a bulge is located in the middle of the sequences, a ≥ 2 nt middle bulge was observed to significantly reduce the folding probability and increase the unfolded or misfolded fractions. This is probably because both the fully-folded and intermediate states of middle bulged G4s would be expected to destabilzed by bulgeclosure entropies. It should be noted that the unfolded or misfolded fractions may include mechanically weak confor-  T2B-2  T3B-2  T5B-2  T7B-2   T5B-3  T3B-3  TB-3 T2B-3 T7B-3  < 40 pN). mations (such as two-quartet G4s) which could unfold at low force (< 4 pN) as reported by previous fluorescenceforce spectroscopy measurements (35). At low forces, the end-to-end extension of ssDNA (∼16 nt) and misfolded G4s are similar (unfolding step sizes < 3 nm), thus cannot be directly detected by force-based measurements (43). The mechanically weak conformations hindered the formation of mechanically stable G4 structures, thereby reducing the folding probability, folding rate, and thermal stability of G4s.