*3.1. GAG Binding Potential of MMP-8*

To assess whether MMP-8 is likely to interact with GAGs, we evaluated its electrostatic surface potential (ESP), as discussed in the literature (Figure 2A) [43]. The ESP map of MMP-8 revealed a catalytic site surrounded by electropositive residues that may serve as GAG or NSGM interacting residues. In fact, multiple electropositive microdomains are located close by, which could possibly favor binding to dimeric and/or trimeric GAG-like NSGM molecules. Thus, we reasoned that there is a high probability of identifying an NSGM that inhibits MMP-8.

Interestingly, we noted that the S<sup>1</sup> ′ pocket of MMP-8 is fairly electropositive in nature. The S<sup>1</sup> ′ pocket is the most varied among the different MMPs. In fact, MMPs can be grouped in the order of the depth of their S<sup>1</sup> ′ pockets. MMPs have either "shallow", "intermediate", or "deep" S<sup>1</sup> ′ pockets [3,44]. MMP-8 is categorized as one with an "intermediate" S<sup>1</sup> ′ pocket. Apart from the size of the S<sup>1</sup> ′ pocket, the residues in the S<sup>1</sup> ′ specificity loop are also considerably different among different MMPs [45]. These differences in the S<sup>1</sup> ′ pocket and its specificity loop are currently being exploited to develop selective MMP inhibitors [44]. Thus, to assess if the electropositive nature of the S<sup>1</sup> ′ pocket is common among MMPs in the "intermediate" category, we analyzed the ESPs of MMP-2 and -9, two other "intermediate" S<sup>1</sup> ′ MMPs [3]. Surprisingly, both MMP-2 and -9 displayed an electronegative S<sup>1</sup> ′ pocket in comparison with that of MMP-8 (Figure 2B,C). Considering these differences in the S<sup>1</sup> ′ pocket of MMP-8 with closely related MMPs, we predicted that the inhibition of MMP-8 by GAGs or NSGMs that target the S<sup>1</sup> ′ pocket is likely to yield a good selectivity.

−

′

′

**Figure 2.** Electrostatic surface potential (ESP) map of MMP-8 (**A**), MMP-9 (**B**), and MMP-2 (**C**) showing electropositive regions (colored blue) that may serve as the site of binding for GAGs and NSGMs. Catalytic Zn2<sup>+</sup> is shown as a yellow sphere. Note the differences in the nature of ESPs between closely related MMPs, particularly inside the S<sup>1</sup> ′ pocket (blue vs. red), despite their relatively good sequence similarity/homology.

### *− 3.2. Structure*−*Activity Relationships for the Library of NSGMs*

To test our hypotheses, we screened our NSGM library against MMP-8 using a fluorogenic substrate hydrolysis assay, as described previously [20]. The screening of the NSGM library resulted in a wide range of inhibitor efficacies (0–100%) against MMP-8 (Figure 3). Interestingly, the inhibition appeared to be related to certain sulfated scaffolds with a majority, e.g., apigenins, catechin, glucoside, inositol, luteolin, morin, phloretin, quercetins, and resveratrol, showing minimal inhibition of MMP-8. This is unusual, because one would expect that the highly sulfated agents, e.g., those carrying more than six sulfate groups, could be expected to target electropositive surfaces more easily. Yet, the results imply that the electropositive regions on MMP-8 are very discriminatory.

**Figure 3.** Residual activity of MMP-8 (10 nM) in the presence of each NSGM (100 µM) in Tris-buffered saline containing 10 mM CaCl<sup>2</sup> , 1 µM ZnCl<sup>2</sup> , pH 7.5 for 10 min at 37 ◦C. The residual MMP-8 activity was measured using a fluorogenic substrate. All of the measurements were performed at least in duplicate. Error bars represent ± 1 standard deviation (SD).

Only five NSGMs (**26**, **38**, **40**, **41**, and **42**) demonstrated an excellent MMP-8 inhibition (>80%). These NSGMs belong to the sulfated benzofuran and sulfated quinazolinone scaffolds. When profiled for the quantitative measurement of MMP-8 potency (Figure 4), the five MMP-8 inhibitors were found to exhibit a three-fold range of potency (*IC*<sup>50</sup> = 11–34 µM, Table 2).

**26 38 40 41 42**


**log[NSGM] mM**

**Residual MMP-8 activity (%)**

**Δ**

0

20

40

60

**Resdiual MMP-8 activity (%)**

80

100

120

**1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58**

**NSGMs chosen for IC<sup>50</sup> studies**

**NSGMs**

− **Figure 4.** Direct MMP-8 inhibition profiles of the most promising NSGMs, including **26**, **38**, **40**, **41**, and **42**, at pH 7.5 and 37 ◦C. Solid lines represent the data fitted to the standard sigmoidal dose−response equation to derive IC<sup>50</sup> and ∆Y, which refer to the potency and efficacy of inhibition, respectively.

**Table 2.** Direct inhibition of MMP-8 by selected NSGMs.


*<sup>a</sup>* Obtained by regression analysis of the dose-dependence of the MMP-8 activity in Tris-buffered saline containing 10 mM CaCl2, 1 µM ZnCl2, pH 7.5, at 37 ◦C. IC<sup>50</sup> and ∆Y refer to potency and efficacy of inhibition, respectively. *<sup>b</sup>* Error refers to ± 1 SD.

At the level of the MMP-8–GAG system, these results are interesting on several fronts. First, the inhibition profiles support the hypothesis that the family of NSGMs is likely to yield an MMP-8 inhibitor, given its structural diversity. The five NSGMs could be classified as "hits" with moderate potency, which will require secondary drug design efforts to yield "lead(s)". Second, the hit yield, i.e., only 5 out of the 58 studied, is relatively low, which suggests excellent weeding-out by MMP-8. Third, not all NSGMs inhibit MMP-8 fully (∆Y > 90%). The lone sulfated benzofuran **26** displays a partial inhibition profile (∆Y = 75%), which presents the possibility of regulating the MMP-8 activity (∆Y = 20–80%), rather than knocking it out completely. Recently, small molecule regulation of soluble enzymes has been discovered and was found to offer beneficial properties [38,46,47]. Similarly, regulation of the MMP-8–NSGM **26** system may offer the benefit of retaining basal levels of the MMP-8 activity that is critical for optimal growth. Fourth, the result that none of the highly sulfated mimetics (NSGMs **1**–**14**, **57**, or **58**), which carry more than six sulfate groups, were active against MMP-8 (Figure 3) implies significant contributions of the aromatic groups. This supports the hypothesis on dual ionic and non-ionic forces governing recognition.

Among the sulfated benzofurans, the closely related monomers (**15**–**25**) and dimers (**27**–**35**) did not inhibit MMP-8 in contrast to **26**, which was one of the three most active NSGMs. In each case, the structural changes were primarily in the terminal substituents (R<sup>1</sup> and R4, Figure 1), suggesting a stringent size dependence. Among the sulfated quinazolinones **36**–**42**, agents with linkers of less than six carbons did not inhibit MMP-8. This implies that the NSGM binding site on MMP-8 is likely to contain two sub-sites spaced several angstroms apart.
