GRUENDER, Circuit Judge.
Mary Ann Ramos and her son, Earl James Ramos, were convicted on several drug-distribution counts. Mary Ramos argues that the district court1 erred by denying her motion for judgment of acquittal. Both defendants contend that the court improperly calculated their advisory sentencing guidelines range. We affirm.
I.
Mary Ramos managed an iWireless store in Cedar Rapids, Iowa. On May 28, 2013, two Drug Enforcement Administration (“DEA”) agents posed as customers and entered her store. When the agents asked Mary if she sold “potpourri,” Mary nodded and retrieved packets labeled “Mr. Happy” and “Mr. Nice Guy.” The agents selected cotton-candy-flavored “Mr. Nice Guy.” While walking to the register, Mary asked if the agents needed rolling papers. They declined and then paid $26.75 for the 10.2 gram packet. Although the “Mr. Nice Guy” packet bore a label indicating that its contents were “100% Cannabinoid Free/ DEA Compliant,” later testing at a DEA laboratory revealed that the packet contained organic plant material sprayed with the Schedule I controlled substance XLR-11, a synthetic cannabinoid.
Several weeks later, a confidential informant working with the Tri-County Drug Enforcement Task Force called Mary and asked to meet at 9:50 p.m. When Mary asked what the caller was “trying to get,” the informant requested “Mr. Nice Guy” and “whatever jar you got.” Mary said that she had “Blue” but did not have “Mr. Nice Guy.” Mary instead offered the caller “Mr. Happy” and “Insane,” which were available in quanti[913]*913ties of twelve and ten grams, respectively. The informant asked for “Mr. Happy,” and he stated that he wished to spend “around 80 or 90” on “the bath salt” and the “Mr. Happy.” Mary drove to meet the informant at a gas station in a nearby town and sold the informant one packet of “Mr. Happy” and one jar of “Blue” for a total of $75. Mary did not charge the informant tax for the purchase nor did she later process the transaction through her register at the iWireless store. A later DEA test determined that the “Blue” weighed 0.2 grams and contained pyrrolidinopentiophene (á-PVP), a substance with a chemical structure substantially similar to the Schedule I controlled substance methylenedioxypyrovalerone (“MDPV’) and with a pharmacological effect substantially similar to MDPV, cocaine, and methamphetamine.
Not long after these incidents, the DEA and Cedar Rapids police executed a search warrant at Mary’s store. The agents and officers recovered “Blue” from a drawer under the register counter. They also located hundreds of packets containing synthetic cannabinoids around the store, including in a drawer under the counter, a back storage room, and the back office. None of the synthetic cannabinoid products were advertised in the store. Nearly all of these products contained XLR-11, and some also contained UR-144, a second synthetic cannabinoid. Labels on many of these packets noted that the product should not be consumed by humans. The agents and officers also found smoking paraphernalia, including glass pipes and rolling papers. They found no loose tobacco in the store.
The DEA and Cedar Rapids police also searched Mary’s car and home. In the car, the agents and officers found an unloaded handgun and four boxes of ammunition immediately next to a box containing several containers of “Blue.” They also found packets containing synthetic cannabinoids in the back pocket of the driver’s seat. In Mary’s home, the officers and agents found more synthetic cannabinoids and another box of “Blue.” The box’s label read “scouring powder” and bore the image of a silhouetted woman in front of a disco ball. Each jar of “Blue” contained between 0.2 and 0.4 grams of á-PVP.
Mary was indicted for several drug-related counts — including distribution of a controlled substance (XLR-11), see 21 U.S.C. § 841(a)(1), distribution of a controlled substance analogue (á-PVP), see 21 U.S.C. §§ 813, 841(a)(1), possession with intent to distribute a controlled substance (XLR-11), see 21 U.S.C. § 841(a)(1), possession with intent to distribute a controlled substance analogue (á-PVP), see 21 U.S.C. §§ 813, 841(a)(1) — and possession of a firearm in furtherance of a drug trafficking crime, see 18 U.S.C. § 924(c). At trial, the Government provided evidence regarding Mary’s encounters with the undercover officers and the confidential informant. The Government also called expert witnesses to testify regarding the synthetic cannabinoids and á-PVP. Another Government witness testified regarding the street names of the drugs at issue.
At the conclusion of the trial, the district court instructed the jury. Regarding the controlled substance counts for XLR-11, the court explained that the Government had to prove that Mary “knew that the substance was some kind of prohibited drug.” For the controlled substance analogue counts involving “Blue,” the court instructed the jury:
[T]he government must prove: (1) the defendant knew that Alpha-PVP was intended for human consumption; and (2) the defendant knew (a) the chemical structure of Alpha-PVP is substantially similar to the chemical structure of a controlled substance in Schedule I or II; [914]*914and (b) Alpha-PVP either has or was represented by the defendant to have a stimulant, depressant or hallucinogenic effect on the central nervous system that is substantially similar to or greater than the stimulant, depressant or hallucinogenic effect on the central nervous system as a controlled substance in Schedule I or II.
After the jury returned its guilty verdicts on the drug-related counts,2 Mary renewed her motion for judgment of acquittal. She argued that the Government had failed to offer sufficient evidence regarding her knowledge. The district court denied this motion.
Earl Ramos managed the Five Star Snacks and Iowa Wireless store in Waterloo, Iowa. Beginning in 2012, Earl sold synthetic cannabinoids — including XLR-11, UR-144, AM-2201, and JWH-081-and synthetic cathinones from his store. Earl elected not to go to trial and instead pleaded guilty to one count of distributing pen-tedrone, a controlled substance analogue, see 21 U.S.C. § 841(a)(1). The court accepted his plea.
Before sentencing Mary and Earl, the district court conducted a joint evidentiary hearing on the nature of the various synthetic cannabinoids sold by the defendants. Synthetic cannabinoids are Schedule I substances; however, they are not listed in the Guidelines Manual drug-equivalency tables. The court thus sought to determine whether the synthetic cannabinoids were more closely related to pure tetrahy-drocannabinol (“THC”) or marijuana, a plant that naturally contains THC. This decision guided the court’s conclusion about which marijuana-equivalency ratio to use when calculating the defendants’ base offense levels under the sentencing guidelines. USSG § 2D1.1 cmt. 6. If the synthetic cannabinoids were more closely related to THC, the Guidelines Manual dictated a 1:167 marijuana-equivalency ratio, meaning that one gram of the synthetic cannabinoid product would be treated as 167 grams of marijuana. USSG § 2D1.1 cmt. 8(D). In contrast, if marijuana were the more similar controlled substance, the applicable ratio would be 1:1. Id.
At the hearing, the Government called Dr. Jordan Trecki, a DEA pharmacologist, to testify regarding the pharmacological effect and potency of the synthetic canna-binoids. Dr. Trecki explained that each of the synthetic cannabinoids -involved in the case has an effect on the central nervous system that is substantially similar to THC. He testified that all but two of the synthetic cannabinoids sold by the defendants are at least as potent as, if not more potent than, THC. Dr. Trecki also discussed the adverse side effects of various synthetic cannabinoids, including hallucinations, psychoses, severe agitation, and excited delirium. Finally, Dr. Trecki noted that synthetic cannabinoids are distinguishable from marijuana in several ways, including (1) the lack of a chemical moderating the effects of THC that is present in marijuana and (2) the synthetic drugs’ increased likelihood of producing seizures, coma, and death. Based on this testimony, the court concluded that THC was the scheduled substance most closely related to the synthetic cannabinoids at issue. The court thus applied a 1:167 marijuana-equivalency ratio to determine the defendants’ base offense levels and advisory sentencing guidelines ranges. The court ultimately sentenced Earl to 57 months’ imprisonment, a sentence within his sentencing guidelines range, and Mary to 60 months’ imprisonment, a downward variance from her advisory guidelines range of 97-121 months.
[915]*915II.
The defendants raise several issues on appeal. Mary contends that the district court erred by denying her motion for judgment of acquittal because the evidence was insufficient to support her conviction. In addition, both Mary and Earl argue that the court improperly calculated their sentencing guidelines ranges based on the erroneous finding that THC was more closely related to the synthetic cannabi-noids at issue.
A.
Our court reviews de novo the denial of a motion for judgment of acquittal. United States v. Malloy, 614 F.3d 852, 860 (8th Cir.2010). “In doing so, we ‘view the evidence in the light most favor-' able to the government, resolving eviden-tiary conflicts in favor of the government, and accepting all reasonable inferences drawn from the evidence that support the jury’s verdict.’ ” Id. at 860-61 (alteration omitted) (quoting United States v. Thomas, 565 F.3d 438, 441 (8th Cir.2009)). We reverse “only if no reasonable jury could have found the defendant guilty beyond a reasonable doubt.” United States v. Brooks, 715 F.3d 1069, 1081 (8th Cir.2013) (quoting United States v. Espinosa, 585 F.3d 418, 423 (8th Cir.2009)). Our court applies this strict standard “even when the conviction rests entirely on circumstantial evidence.” United States v. Tillman, 765 F.3d 831, 833 (8th Cir.2014) (quoting United States v. Worman, 622 F.3d 969, 977 (8th Cir.2010)).
Knowledge is one element of drug-distribution charges under 18 U.S.C. § 841(a). See United States v. Nichols, 808 F.2d 660, 663 (8th Cir.1987). Mary contends on appeal that the Government failed to prove that she knew the synthetic cannabinoid products contained XLR-11 or some other controlled substance. The Government, however, was “not required to prove that the defendant actually knew the exact nature of the substance with which [s]he was dealing.” United States v. Sheppard, 219 F.3d 766, 769 (8th Cir.2000) (quoting United States v. Jewell, 532 F.2d 697, 698 (9th Cir.1976) (en banc)). “[T]he ‘knowingly’ element of th[e] offense refers to a general criminal intent, i.e., awareness that the substance possessed was a controlled substance of some kind.” United States v. Noibi, 780 F.2d 1419, 1421 (8th Cir.1986). “Since the factfinder can seldom know with certainty what someone actually knows, knowledge must necessarily be shown circumstantially.” Id.
During trial, the Government presented evidence tending to show that Mary knew she was dealing with illegal drugs when she sold the XLR-11. Though Mary openly sold glass pipes and other smoking paraphernalia, the packets containing synthetic cannabinoid products were not on public display in the store nor advertised in any way. One undercover officer who took part in the controlled purchase testified that Mary produced packets of “Mr. Happy” and “Mr. Nice Guy” only after the agent requested “potpourri,” a street name for synthetic cannabinoid products. The label on the packets of “Mr. Happy” and “Mr. Nice Guy” stated that the product was not for human consumption; however, the “Mr. Nice Guy” packet advertised a cotton-candy flavor. And Mary, unprompted, asked if the officers needed rolling papers with the “Mr. Nice Guy” purchase. After the agents declined, Mary charged $25 for the small, ten-gram packet of purported “potpourri.” Finally, the jury heard that the search of Mary’s store, car, and home yielded hundreds of packets containing synthetic cannabinoids stored out of public view. Viewed in the light most favorable to the jury’s verdict, this evidence was sufficient to demonstrate Mary’s awareness that she possessed and sold a controlled substance of some kind.
[916]*916Mary also contends that the Government failed to meet its burden to prove her knowledge with respect to á-PVP, the controlled substance analogue in “Blue.” She argues that the jury instructions created a heightened knowledge requirement and that her conviction cannot stand because the Government failed to meet its resultant burden. As an initial matter, Mary is incorrect in her assertion that we should assess sufficiency under the instructions given to the jury. The Supreme Court recently held that “a sufficiency challenge should be assessed against the elements of the charged crime, not against the erroneously heightened jury instructions.” Musacchio v. United States, 577 U.S. -, 136 S.Ct. 709, 714, 193 L.Ed.2d 639 (2016). In any event, it is not clear that the instructions in the present case laid out a heightened knowledge element; instead, they described one of two possible methods of proving knowledge in analogue cases.
The Supreme Court clarified the two ways the Government may prove knowledge in analogue cases last term in McFadden v. United States, 576 U.S. -, 135 S.Ct. 2298, 192 L.Ed.2d 260 (2015). The Court explained that the Government either must (1) show the defendant knew she was dealing with some controlled substance, regardless of whether she knew the identity of the substance, or (2) show that the defendant knew the specific features of the substance that make it a controlled substance analogue. Id. at 2302. These features include (a) having a “ ‘chemical structure ... which is substantially similar ‘to the chemical structure of a controlled substance in schedule I or II’ ” and (b) having “ ‘a stimulant, depressant, or hallucinogenic effect on the central nervous system that is substantially similar to or greater than’ the effect of a controlled substance in schedule I or II ... or is represented or intended to have that effect with respect to a particular person.” Id. at 2305 (quoting 21 U.S.C. § 802(32)(A)). Such knowledge suffices because “[a] defendant who possesses a substance with knowledge of those features knows all of the facts that make his conduct illegal, just as a defendant who knows he possesses heroin knows all of the facts that make his conduct illegal.” Id.
The court’s instructions at Mary’s trial closely matched the second method of proving knowledge under McFadden,3 The court instructed the jury that the Government had to prove (1) that Mary knew that á-PVP was intended for human consumption4 and (2) that she knew (a) that the chemical structure of á-PVP is substantially similar to the chemical structure of a controlled substance in Schedule I or II and (b) that á-PVP either has or was represented by Mary to have a stimulant, depressant, or hallucinogenic effect on the central nervous system that is substantially similar to or greater than the stimulant, depressant, or hallucinogenic effect on the central nervous system as a controlled substance in Schedule I or II.5
[917]*917Mary contends that the Government did not prove the knowledge element of the charged crime because the Government failed to offer direct evidence regarding her knowledge of the chemical structure and pharmacological effects of á-PVP.6 However, as with prosecutions involving substances listed on the drug schedules, the Government did not need to present direct evidence of Mary’s knowledge of the analogue’s nature and effects. Instead, the Government could satisfy its burden through circumstantial evidence of knowledge. See United States v. Carlson, 810 F.3d 544, 552 (8th Cir.2016). Examples of such circumstantial evidence in analogue cases include “a defendant’s concealment of his activities, evasive behavior with respect to law enforcement, knowledge that a particular substance produces a ‘high’ similar to that produced by controlled substances, and knowledge that a particular substance is subject to seizure at customs.” McFadden, 135 S.Ct. at 2304 n. 1.
Judged under this standard, the evidence was sufficient to show that Mary knew the specific features of á-PVP that make it a controlled substance. First, the evidence suggested that she knew she was in possession of an analogue to a controlled substance with a chemical structure similar to a drug listed on the federal drug schedules. The jury heard a recorded conversation in which the confidential informant expressed his desire to buy bath salts. A Government witness at trial explained that this term is a street name for “generic forms of meth or cocaine.” And the evidence makes clear that Mary understood this meaning. Cf. Sullivan, 714 F.3d at 1107 (concluding that a defendant’s knowledge that “bath powder” was illegal supported a reasonable inference that he knew powder he possessed and described as “bath powder” contained a controlled substance analogue); Carlson, 810 F.3d at 552 (noting that knowledge of an analogue’s effects may be probative of the defendant’s knowledge of chemical structure). The cloak-and-dagger nature of the transaction with the confidential informant evidenced an attempt to conceal the sale. The jury heard that Mary met the confidential informant to conduct the sale at a gas station just before 10:00 p.m. Mary did not charge tax or provide a receipt for the purchase, nor did she later process the transaction through her register at the iWireless store. In addition, though the jars of “Blue” recovered in this case were labeled “scouring powder,” the jury heard that Mary charged the informant $50 for 0.2 grams. This price far exceeds the cost of comparable cleaning products; indeed, a Government witness testified that twenty to twenty-four ounces — roughly 567-680 grams — of Comet or a similar powder would cost less than a dollar.
Second, the circumstantial evidence supported the inference that Mary understood the pharmacological effects of “Blue,” specifically that it produced a high similar to a controlled drug. The boxes of “Blue” in Mary’s possession bore an image of a silhouetted woman in front of a disco ball, suggesting the substance’s use as a party drug. During the transaction, the informant explained to Mary that he had been visiting another location to make similar purchases multiple times per day. Finally, Mary stated that the informant would be [918]*918“real happy” following the late-night transaction. Viewed in the light most favorable to the verdict, this evidence allowed the jury to infer Mary’s knowledge of the specific features of the substance that make it a controlled substance analogue, that is, its use, chemical nature, and pharmacological effects. Accordingly, the evidence was sufficient under the second method of proving knowledge outlined in McFadden. We thus affirm the district court’s denial of the motion for judgment of acquittal.
B.
Mary and Earl Ramos also appeal the calculation of their advisory sentencing guidelines ranges, arguing that the district court improperly determined their base offense levels when it found that the synthetic cannabinoids sold by the Ramoses were more like pure THC than an equal quantity of marijuana. We join our sister circuits in treating this similar-substance question as a factual determination. See United States v. Malone, 809 F.3d 251, 257 (5th Cir.2015); United States v. Brey, No. 15-10165, 627 Fed.Appx. 775, 2015 WL 5521181, at *4 (11th Cir. Sept. 21, 2015) (unpublished); United States v. Beckley, 515 Fed.Appx. 373, 375 (6th Cir.2013) (unpublished); United States v. Chowdhury, 639 F.3d 583, 585 (2d Cir.2011). We thus review for clear error. See United States v. Stanley, 362 F.3d 509, 511 (8th Cir.2004).
The synthetic cannabinoids at issue in this case, such as XLR-11, are Schedule I controlled substances, see 21 U.S.C. § 812 Schedule 1(d); however, they are not referenced in the guidelines drug-equivalency tables. In cases involving controlled substances not specifically referenced in the guidelines, the district court must use the marijuana-equivalency ratio for the most closely related controlled substance found in the drug-equivalency tables. USSG § 2D1.1, cmt. 6. To identify the most closely related substance, the court must consider, to the extent practicable, three factors:
(A) Whether the controlled substance not referenced in this guideline has a chemical structure that is substantially similar- to a controlled substance referenced in this guideline.
(B) Whether the controlled substance not referenced in this guideline has a stimulant, depressant, or hallucinogenic effect on the central nervous system that is substantially similar to the stimulant, depressant, or hallucinogenic effect on the central nervous system of a controlled substance referenced in this guideline.
(C) Whether a lesser or greater quantity of the controlled substance not referenced in this guideline is needed to produce a substantially similar effect on the central nervous system as a controlled substance referenced in this guideline.
USSG § 2D1.1, cmt. 6. The parties acknowledge that the first factor did not aid the district court’s analysis; synthetic can-nabinoids do not have a chemical structure similar to either THC or marijuana. Accordingly, our review focuses on the district court’s findings with respect to factors B and C.
During the evidentiary hearing, Dr. Trecki testified in detail about facts relevant to factor B. He explained that the synthetic cannabinoids affect the central nervous system in a manner substantially similar to THC. Dr. Trecki testified to the synthetic cannabinoids’ euphoric effect and their potential to cause hallucinations, psychoses, and severe agitation. He also noted that both the synthetic cannabinoids and THC cause heart palpitations and reddening of the conjunctiva. Although Dr. Trecki noted that marijuana also produces some of these effects because marijuana contains THC, he explained that marijuana plants also contain a moderating chemical, cannabidiol, that mediates THC’s effects [919]*919and reduces euphoria. Finally, Dr. Trecki explained that synthetic cannabinoids carry risks that are “very uncommon for marijuana,” including seizures, coma, and death.
Dr. Trecki also provided testimony relevant to the district court’s analysis under factor C. He explained that equal amounts of pure synthetic cannabinoids and pure THC produce similar effects on the central nervous system. Indeed, discrimination studies showed that animals could not differentiate between XLR-11, one of the synthetic cannabinoids, and THC. He further explained that many of the synthetic cannabinoids sold by the Ramoses were more potent than THC. When asked about the medium bearing the synthetic cannabi-noids in the Ramoses’ cases — i.e., the potpourri — Dr. Trecki noted that spraying synthetic cannabinoids on organic plant material would not change the active substance’s nature, character, or potency. The process simply would dilute the synthetic cannabinoids over a greater volume of material. In light of this testimony, the court concluded that THC was the most similar substance to the synthetic cannabi-noids sold by the defendants. Accordingly, the court calculated the base offense levels using a 1:167 marijuana-equivalency ratio provided for THC. See USSG § 2D1.1 n.8(D). We see no clear error in the court’s determination. See Malone, 809 F.3d at 258 (affirming the district court’s conclusion that AM-2201, a synthetic cannabinoid, was most closely related to pure THC); cf. Carlson, 810 F.3d at 556 (rejecting the defendant’s legal challenge to the application of a 1:167 marijuana-equivalency ratio for synthetic cannabi-noids, such as XLR-11, UR-144, and AM-2201).
The dissent disagrees with this conclusion and instead adopts the Ramoses’ position that the district court improperly applied factor C. The dissent contends that the court should have examined the effects of “synthetic cannabinoid potpourri,” the mixture of inert plant material and synthetic cannabinoids, rather than pure synthetic cannabinoids alone. Infra 921. However, this argument runs afoul of the plain text of the Guidelines Manual. When cases involve controlled substances not specifically referenced in the drug-equivalency tables, Comment 6 to § 2D1.1 tells the court to compare “the controlled substance[s] not referenced in th[e] guideline” to the substances listed in the drug-equivalency tables. Here, the record makes clear that “the controlled sub-stancefs]” at issue were the synthetic can-nabinoids, The term “controlled substance” is defined by statute as “a drug or other substance, or immediate precursor, included in schedule I, II, III, IV, or V.” 21 U.S.C. § 802(6). And while synthetic cannabinoids, such as XLR-11, are listed in Schedule I, see 21 U.S.C. § 812 Schedule 1(d), “synthetic cannabinoid potpourri” is not independently listed on any drug schedule. Mary was found guilty of distributing the Schedule I controlled substance XLR-11. Earl’s plea agreement identified the Schedule I substances at issue as the synthetic cannabinoids. In light of these facts, we conclude that the district court properly applied the plain language of factor C when it compared only the effect of “the controlled substances” — the synthetic cannabinoids — to the substances listed in the drug-equivalency tables. See United States v. Figueroa, 647 F.3d 466, 470 (2d Cir.2011) (remanding for reconsideration when the district court failed to apply the three-factor test from § 2D1.1 and instead relied on the synthetic drug’s street-use counterpart to determine the most closely related listed substance).7
[920]*920With this principle in mind, the Ramoses’ arguments challenging the district court’s base offense level calculations lack merit. After identifying THC as the most similar substance, the court correctly calculated the Ramoses’ base offense levels using the entire weight of the substances containing the synthetic cannabinoids. Application Note A to § 2D1.1 states: “Unless otherwise specified, the weight of a controlled substance set forth in the table refers to the entire weight of any mixture or substance containing a detectable amount of the controlled substance.” Stated another way, “drug quantity under the guidelines includes the entire weight or volume of any mixture containing a detectable amount of controlled substance, without regard to purity or concentration.” United States v. Stewart, 761 F.3d 993, 1001 (9th Cir.2014). Only if a user must separate the other materials from the controlled substance prior to use should the court calculate the drug quantity without including this weight. USSG § 2D1.1 cmt. I.8 And only if the Guidelines Manual carves out an express exception does the mixture-or-substance rule not apply. See, e.g., § 2D1.1 & n.(B), (G); see also United States v. Shabazz, 933 F.2d 1029, 1032 (D.C.Cir.1991). Here, the controlled substances were sold in the form of a mixture: synthetic cannabinoids sprayed onto inert plant material. See Chapman v. United States, 500 U.S. 453, 461-62, 111 S.Ct. 1919, 114 L.Ed.2d 524 (1991) (discussing mixtures). This plant material was not removed prior to the drug’s use, nor was it easily separable. Likewise, synthetic can-nabinoids are not among the listed exceptions to the mixture-or-substance rule. Accordingly, the district court properly found that the weight of the mixture dictated the drug-quantity weight at sentencing.
In sum, because the court correctly applied the guidelines and because Dr. Trecki’s testimony supported the court’s factual conclusion that THC was the most similar substance to the synthetic cannabi-noids at issue, we find no clear error in the application of the 1:167 marijuana-equivalency ratio and, therefore, the determination of the Ramoses’ base offense levels. In the absence of any other challenges, we uphold the Ramoses’ sentences.
For the foregoing reasons, we affirm Mary’s conviction and sentence and Earl’s sentence.